Rethinking p16, p53, and HPV in HNCSCC through lessons from glioblastoma subclonal evolution toward patient-centric N-of-1 single-cell RNA sequencing paradigm

Abstract

Head and neck cutaneous squamous cell carcinoma (HNCSCC) remains underexplored compared to oropharyngeal squamous cell carcinoma, particularly in relation to human papillomavirus (HPV) and molecular markers such as p16 and p53. While p16 is a well-established surrogate for HPV in oropharyngeal cancer, our review highlights its unreliable role in HNCSCC, where positivity is instead associated with recurrence and metastasis. Similarly, p53 illustrates a dual role - wild-type as a genomic safeguard, mutated as an oncogenic driver - complicating prognostication. Methodological considerations, including the limitations of immunohistochemistry for HPV detection, underscore the need for multi-method and molecular validation in future studies. Ultraviolet radiation is posited as a key modifier of p16 function, decoupling expression from tumor suppression. To contextualize these findings, we draw parallels to glioblastoma (GBM), where subclonal evolution, p53 dysfunction, and intratumoral heterogeneity drive relapse despite aggressive multimodal therapies. GBM exemplifies how bulk-level biomarker generalizations often obscure dynamic cellular ecosystems, reinforcing the necessity of single-cell and spatial approaches. Multi-omics integration - encompassing genome, transcriptome, proteome, and tumor microenvironment mapping - coupled with single-cell RNA sequencing and spatial transcriptomics, offers a path forward for resolving subclonal dynamics in both HNCSCC and GBM. These technologies provide the resolution needed to track tumor-immune-stromal co-evolution, identify therapy-resistant clones, and anticipate recurrence. We argue for a N-of-1, patient- and cell-centric paradigm that reframes biomarkers not as static surrogates but as dynamic readouts of cancer evolution across time and tissue contexts. Conceptually, we propose kinetic and microenvironmental frameworks (e.g., “load-and-lock” barriers; dormancy and immune-synapse stabilization) as hypothesis-generating avenues to stall clonal handoffs and improve outcome prediction. Together, these perspectives argue for revised biomarker frameworks in HNCSCC and ethically inclusive, mechanism-anchored studies that bridge discovery with individualized care. By bridging insights from HNCSCC with the lessons of GBM, this review underscores the need for ethically inclusive, mechanistically informed frameworks that integrate subclonal evolution, biomarker re-interpretation, and precision-personalized hybrid models. Such an approach will be essential for advancing from one-size-fits-all strategies to individualized lifetime cancer care.

Key Words: p16; p53; Human papilloma virus; Head and neck cutaneous squamous cell carcinoma; Glioblastoma; Ultraviolet radiation; Subclonal evolution; Single-cell RNA sequencing; Spatial transcriptomics; N-of-1 paradigm

Core Tip: Nam et al’s article provides a compelling investigation into the mysterious interaction between human papillomavirus, p16, and p53 in Asian populations. This investigation is part of the investigation into the pathogenesis and prognosis of cancer. The results of this study not only contribute to our expansion of knowledge regarding head and neck cutaneous squamous cell carcinoma, but they also question our preconceived assumptions regarding biomarkers that have traditionally been associated with positive outcomes.

INTRODUCTION

Nam et al’s article[1] provides a compelling investigation into the mysterious interaction between human papillomavirus (HPV), p16, and p53 in Asian populations (Figure 1). This investigation is part of the investigation into the pathogenesis and prognosis of cancer in an ethnic group. Head and neck cutaneous squamous cell carcinoma (HNCSCC) is a complex and underexplored variant of squamous cell carcinoma (SCC), particularly concerning its association with HPV and molecular markers like p16 and p53. This retrospective cohort study analyzed 62 Asian patients with HNCSCC to assess the prevalence of HPV infection and the prognostic implications of p16 and p53 expression. While HPV plays a well-established role in certain head and neck cancers, its significance in HNCSCC remains uncertain. Notably, this study found that p16 expression is not a reliable surrogate marker for HPV in HNCSCC, in contrast to its role in oropharyngeal (OP) SCC, where it is typically associated with better prognosis. Instead, p16 positivity in HNCSCC was linked to higher recurrence and metastasis rates, raising questions about the underlying biological mechanisms driving this discrepancy.

Figure 1 This diagram illustrates the relationship between human papillomavirus, p16 expression, p53 mutation, and their roles in head and neck squamous cell carcinoma progression and prognosis. Human papillomavirus infection, p16 expression, and p53 mutations are interconnected factors influencing head and neck squamous cell carcinoma (HNCSCC) development, progression, and prognosis. While p16 is favorable in oropharyngeal squamous cell carcinoma, it is less reliable in HNCSCC, especially when combined with p53 mutations, which worsen outcomes. These biomarkers are essential for understanding and managing HNCSCC. HPV: Human papillomavirus; HNCSCC: Head and neck squamous cell carcinoma; SCC: Squamous cell carcinoma.

Key findings and implications include (Figure 1): (1) HPV prevalence and p16 expression: HPV was detected in only 8.06% of HNCSCC cases, significantly lower than in OPSCC. p16 positivity was observed in 38.71% of cases, but its predictive value for HPV was low, suggesting different oncogenic drivers in HNCSCC. The dual role of p16 in SCC: Unlike in OPSCC, where p16 overexpression is a favorable prognostic marker, in HNCSCC, it correlated with higher recurrence (50%) and metastasis (33.33%). The study suggests that ultraviolet (UV) radiation exposure may play a key role in altering p16’s tumor-suppressive functions, but further experimental or epidemiological data are needed to substantiate this hypothesis; (2) p16 positivity and UV-linked mechanisms - while p16 overexpression in OPSCC is often associated with improved prognosis, their findings suggest that in HNCSCC, p16 positivity instead correlates with recurrence and metastasis[1]. A plausible explanation lies in the mutagenic effect of UV radiation on keratinocytes. Chronic UV exposure is known to induce mutations in CDKN2A and RB pathway regulators, potentially overriding the canonical tumor-suppressive checkpoint function of p16[2]. In this context, p16 overexpression may represent a cellular stress response to UV-induced genomic instability, rather than a surrogate for viral oncogenesis. Epidemiological studies of sun-exposed cutaneous SCC support this interpretation, showing that high UV burden is linked with both increased p16 expression and poorer clinical outcomes. Experimental evidence also suggests that UV-driven mutations in downstream cell-cycle regulators can decouple p16 expression from its tumor-suppressive role, converting it into a biomarker of cellular distress rather than protection. For example, UV-driven DNA damage may override p16-mediated checkpoint control as evidenced in advances from preclinical studies, current clinical trials, and clinical applications of chemical inhibitors targeting key DNA damage response proteins - including DNA-dependent protein kinase catalytic subunit, Ataxia telangiectasia mutated (ATM)/ataxia telangiectasia mutated Rad3-related kinase, the Mre11-Rad50-Nbs1 complex, polyadenosine-diphosphate-ribose polymerases, mediator of DNA damage checkpoint protein 1, Wee1-like protein kinase, DNA ligase 4, cyclin-dependent kinase 1, breast cancer suppressor gene 1, checkpoint kinase 1, and hypoxia-inducible factor 1 and highlight challenges for ionizing radiation-induced signal transduction and targeted therapy in light of recent progress in radiobiology[3]. These mechanistic considerations warrant further study in prospective cohorts with detailed UV-exposure histories and molecular profiling; (3) p53 mutations and tumor aggressiveness: P53 mutations were present in 72.58% of cases, reinforcing their role as a critical driver of HNCSCC progression. The study highlights the dual nature of p53 in cancer biology, where its mutations contribute to tumor progression and resistance to apoptosis. Further integration with existing literature on p53’s function in other cancers would strengthen the analysis (refer to “p53 dynamics: A comparison and contrast an example of a double-edged sword: Mutation”); and (4) HPV detection methodology (methods/Limitations section): The study relied on immunohistochemistry (IHC) to detect HPV, despite polymerase chain reaction (PCR) being a more sensitive method. A stronger justification for this choice or alternative detection approaches in future studies would enhance credibility. A limitation of their study is the reliance on IHC for HPV detection, which the authors acknowledge is less sensitive than PCR-based methods. However, IHC remains widely used in routine pathology because it is cost-effective, readily available, and applicable to archival paraffin-embedded tissues, making it pragmatic for retrospective analyses such as ours. Furthermore, IHC provides a morphological context that can be diagnostically valuable. Nevertheless, we concur that future studies should incorporate more sensitive molecular methods, including PCR or in situ hybridization (ISH), ideally in combination, to ensure accurate HPV detection. A multi-method approach would strengthen the validity of HPV prevalence estimates in HNCSCC and allow more definitive conclusions about its prognostic role. The results of this study were HPV positivity (8.06%), p16 expression (38.71%), and p53 expression (72.58%). The finding that there is no link between p16 and HPV not only contributes to our understanding of HNCSCC but also challenges our preconceived notions, prompting a re-evaluation of biomarkers that have traditionally been associated with positive outcomes. Amid ongoing debates surrounding the potential benefits and risks of cancer therapies, bridging these differing viewpoints has become a pressing priority.

A RE-EXAMINATION OF GENERALIZATIONS REGARDING P16 AND HPV IN THE CONTEXT OF SCC

Prior expectations about biomarkers that have historically been linked to favorable prognoses. It is well known that p16 is a surrogate marker for HPV in OPSCC and that its overexpression often indicates a more favorable prognosis. On the other hand, this study’s findings demonstrate a striking contrast: P16 positivity in HNCSCC is associated with increased chances of recurrence and metastasis (Figure 1). This unanticipated finding encourages readers to reevaluate the universality of p16 as a prognostic marker and investigate its nuanced role, which is influenced by factors such as the location of the tumor and the damage caused by UV radiation. Compared to its well-established utility in treating OP malignancies, why does the predictive value of p16 in HNCSCC appear to be lower? According to the study, p16’s tumor-suppressive activity may be disrupted by UV light, which is a hypothesis that is ready for future investigation.

The p16 expression is strongly associated with HPV infection in HNSCC, but it is not a perfect surrogate marker for HPV. While p16’s prognostic value is well-established in OPSCC, its role in non-OPSCC sites such as the oral cavity, larynx, and hypopharynx remains understudied. This study analyzed 80 HNSCC cases and found that p16 expression varied by tumor site (P < 0.001), race (P = 0.031), marital status (P = 0.008), and smoking history (P = 0.014). Patients who were p16 positive had improved survival, particularly in HPV16-positive cases, whereas those who were p16 negative and HPV16 negative had the worst survival outcomes. Future studies with larger, more diverse cohorts and multiple molecular markers are necessary to better define p16’s role as a prognostic indicator in HNSCC. Location-specific, sex-specific, and race-specific incidences of HNSCC are associated with biomarker patterns[4]. Specifically, this group found that (Figure 2): (1) p16 expression and HPV16 correlation: P16 positivity is highly associated with HPV16 infection in HNSCC, especially in OP tumors, but it is not a perfect surrogate marker as it also appears in HPV-negative cases; (2) Site-specific variability: P16 expression differs across tumor sites (P < 0.001), being more prevalent in OP than in non-OP sites (oral cavity, larynx, and hypopharynx) (P < 0.0001), with variations observed between Caucasian Americans and African Americans (P = 0.031), similar to HPV prevalence differences (P = 0.013); (3) Demographic associations: P16 positivity is linked to marital status (P = 0.008) and smoking history (P = 0.014), highlighting socio-behavioral influences on its expression in HNSCC; and (4) Survival outcomes: Patients with p16-positive tumors show improved survival, similar to HPV16-positive cases, while those with p16-negative/HPV16-negative status have the worst survival for all sites combined and for OP specifically. Larger cohorts, including non-OP sites and multiple molecular markers, are essential to fully understand p16’s prognostic significance, especially in distinguishing HPV16-negative p16-positive subgroups.

Figure 2 An illustration depicting Caucasian and African American individuals with a graphical representation of head and neck cancer biomarkers, survival markers, and associated prognosis. The table compares survival rates, hazard ratios, and P values for head and neck cancer patients based on biomarker status and race (Caucasian American vs African American). Specifically, p16+ (better prognosis): P16+ is generally associated with better prognosis, but survival disparities exist between races. p16- (worse prognosis): P16 is linked to significantly worse survival rates, especially among Caucasian patients. Human papillomavirus (HPV) 16+ (better prognosis): HPV16+ is associated with improved survival, though the effect varies between racial groups. HPV16- (worse prognosis): HPV16 results in significantly worse survival outcomes, particularly for Caucasian patients. HPV: Human papillomavirus; CA: Caucasian American; AA: African American; HR: Hazard ratio; CI: Confidence interval; N/A: Not applicable.

P53 DYNAMICS: A COMPARISON AND CONTRAST, AN EXAMPLE OF A DOUBLE-EDGED SWORD: MUTATION

It has been found that p53, which is another molecular marker of SCC, is a consistent factor in the aggressiveness of tumors, in addition to its stem cell root[5]. This study provides more evidence that it has a substantial connection with unfavorable outcomes, establishing it as an important prognostic factor. With its mutation-driven resistance to apoptosis and ability to allow unregulated cell proliferation, p53 provides a story that contrasts the complexity of p16. The data also indicate no substantial correlation between the expression of p53 and the presence of HPV, highlighting the necessity of more extensive molecular research to disentangle these interrelationships.

Nam et al’s analysis[1] reinforces p53 as a consistent marker of tumor aggressiveness in HNCSCC; yet, its role in cancer biology remains unexplored, particularly in the dual nature of p53 in cancer biology, where its mutations contribute to cancer progression vs its normal tumor suppressive functions. In its wild-type form, p53 acts not only as a pivotal tumor suppressor - activated in response to stress signals such as DNA damage to induce cell-cycle arrest, DNA repair, senescence, or apoptosis, but also as a master regulator of genomic stability, orchestrating cell-cycle arrest, DNA repair, and apoptosis - thereby safeguarding cells against malignant transformation[6]. Conversely, TP53 mutations - most of which are missense - abrogate these protected functions[7]. Thus, TP53 mutations not only abolish these protective functions (loss-of-function) but may also confer gain-of-function (GOF) properties - gain oncogenic activities that promote cancer proliferation, invasion, metastasis, metabolic reprogramming, immune evasion, angiogenesis, and therapy resistance[8]. This paradox is well documented across multiple cancers, including lung, breast, and hematologic malignancies, where mutant p53 subverts normal transcriptional programs and drives aggressive phenotypes[9]. By situating Nam et al’s findings[1] within this broader literature, it becomes clear that the unfavorable prognostic impact of p53 expression in HNCSCC reflects the accumulation of mutation-driven alterations rather than the tumor-suppressive activity of wild-type p53.

Such duality highlights the need to stratify patients not only by p53 expression but also by p53 mutation status, functional activity, and potential therapeutic vulnerabilities - a “perspective directly revisiting p53’s context-dependent dual roles (Wilms tumour suppressor vs mutant oncogenic activities still not fully understood, and our discoveries have not yet led to major therapeutic breakthroughs”[10] ever since 1979. First reports identified a 53-54 kDa protein (later named p53) bound to SV40 large T antigen in transformed cells[11]. As an expert in the field, Soussi[12] witnessed and penned the 2010 EMBO reports essay that traces p53’s path from its 1979 discovery and early misclassification as an oncogene to its recognition as a tumor suppressor, using this history to show how prevailing Kuhnian paradigms and experimental artifacts (e.g., mutant constructs) can misdirect a field and why continual re-examination of assumptions is essential[12]. About half of human cancers harbor TP53 mutations, and while GOF theory attributes oncogenic properties to 60%-80% of these, alternative explanations - most notably dominant-negative inhibition of the remaining wild-type p53 before loss of heterozygosity - plus conflicting data keep GOF contested; ultimately, transformation reflects mutation-cell-state interplay, with mutant p53 acting in a consistently context-dependent manner[13]. The “dual nature” concept: The juxtaposition of p53’s canonical tumor suppression and the oncogenic capabilities of its mutants underscores its complex and dualistic roles in cancer biology, suggesting that a therapeutic window[14] based on timely diagnosis is essential in cancer treatment and monitoring cancer-response-driven cancer subclonal evolution at a single cell tracking system of HNSCC[15] based on a spatiotemporal determination of p53 expression dynamic changes with a single-cell platform[16]. All of these debates demand a N-of-1 patient care context for a specific p53 status of the treatment trajectory.

FOSTERING DIALOGUE FOR FUTURE TREATMENT: AN APPEAL FOR ACCURACY IN THE EMERGING METHODS USED TO DETECT AND DIAGNOSE HPV

A multidisciplinary approach is essential to fully capture the complexities of SCC prognosis and treatment. To advance beyond single-marker associations, multi-omics research is needed to illuminate the intricate interplay of molecular and clinical factors. Comprehensive studies of the tumor microenvironment have begun to show how stromal, immune, and vascular components shape SCC progression and therapy response, such as in lung cancer. A six-gene prognostic signature (MYO1E, FEN1, NMI, ZNF506, ALDOA, and MLLT6) was developed from the Cancer Genome Atlas-Lung Adenocarcinoma dataset to stratify patients into high- and low-risk groups[17]; and Lamin B2 as a diagnostic and prognostic biomarker[18]. In parallel, emerging technologies such as single-cell spatial transcriptomics now provide unprecedented resolution of intratumoral heterogeneity, enabling in situ mapping of clonal dynamics and microenvironmental niches[19]. Moreover, single-cell spatial transcriptomics provides a powerful approach to resolve intratumoral heterogeneity at cellular resolution. Recent single-cell and spatial transcriptomic analyses of immune-hot and immune-cold tumors have identified distinct fibroblast subtypes that shape immunological niches and correlate with positive immunotherapy responses[20]. By capturing how tumor cells and their surrounding microenvironment co-evolve, this methodology can yield critical insights that inform novel therapeutic strategies and refine prognostic models. Integration of single-cell RNA sequencing (scRNA-seq) and spatial transcriptomics has demonstrated the therapeutic potential of nitazoxanide in HNSCC, with hub genes identified as independent prognostic factors[21]. Integrating these profiling-omics approaches with clinical data will help delineate how tumor cells and their ecosystems co-evolve under selective pressures such as UV exposure and therapy. Such insights promise to refine prognostic models and reveal novel therapeutic entry points tailored to the unique molecular and cellular landscapes of HNCSCC.

Although the study utilizes IHC for HPV identification, the sensitivity of this method is not nearly as high as that of methods based on PCR. Due to the low HPV positivity rate (8.06%), it is necessary to re-evaluate the analytical approaches to guarantee that complete insights are obtained. The fact that this discrepancy exists highlights the significant role that enhanced detection techniques play in enhancing our comprehension of the prevalence of HPV and its consequences in non-OPSCC. Based on the meta-analysis findings, HPV16 E6 serology is the most sensitive (83.1%) and specific (94.6%) biomarker for HPV-driven OP cancer[22]. However, the accuracy of its performance is influenced by the molecular reference method used. Among the molecular reference methods - HPV DNA, RNA, ISH, and p16 IHC - studies using a combination of at least two methods (HPV DNA, RNA, ISH, or p16 IHC) provided higher specificity compared to using HPV DNA PCR alone. This suggests that a multi-method approach is superior, as relying solely on HPV DNA PCR may lead to lower specificity. If a single method must be chosen, ISH or p16 IHC as stand-alone markers appear to be more reliable than HPV DNA PCR alone. However, given the variability in test performance, combining at least two methods remains the best approach for accurately determining HPV tumor status in OPC cases.

Thus, current pathology commonly reliable ISH diagnosis aligns with findings from a study on oral SCC (OSCC), where p16INK4A IHC and HPV16 E6/E7 mRNA ISH showed a significant correlation with clinical and pathological features of OSCC, particularly in patients with a history of alcohol consumption and tumors located in the hard palate (38% p16INK4A, 36% HPV-16 ISH positivity)[23]. Given this evidence, p16INK4A remains a useful surrogate marker for HPV detection in OSCC, which can be further strengthened by RNA ISH to confirm HPV subtypes and enhance diagnostic accuracy upon taking the profiling-omics in the equation.

EXPANDING THE RANGE OF FOCUS: A CONCENTRATION ON ASIAN POPULATIONS AND BEYOND HNSCC

One of the most remarkable strengths of the study is that it addresses a significant gap in the literature, which has primarily focused on specific cohorts. This approach was achieved by focusing on Asian people, specifically Koreans. With this demographic focus, unique molecular and epidemiological patterns are revealed, and readers worldwide are invited to assess the geographical heterogeneity in cancer biology and the implications of this variation for personalized treatment options. HPV16 variants from Barbadian clinical isolates were found at a unique position in the phylogenetic tree, near the divergence of the European and Asian lineages, whereas Dutch samples clustered within the European branch[24]. Despite the geographical and ethnic differences, sequence analysis revealed few mutational differences, with the locus control region and E5 open reading frame regions key for proper variant classification. Mucosal melanoma incidence varies significantly among ethnic groups, and this study focused on Chinese patients to identify key genetic alterations. Whole-exome sequencing revealed a high copy number variant burden and low single-nucleotide polymorphism burden, with recurrent mutations in FAT atypical cadherin 1 (100% in anorectal cases), phosphoinositol-3-kinase, Kirsten rat sarcoma virus (KRAS), adenomatous polyposis coli, and breast cancer suppressor gene 1, suggesting potential ethnic-specific pathological mechanisms and an association with HPV infection[25]. That long-term retrospective study of 400 HPV-associated HNSCC cases in Southern China found an overall HPV prevalence of 15%, highest in OP cancers (30.7%). HPV-positive cases were more frequently associated with tonsillar invasion, older age, male gender, and strong tobacco/alcohol use, which correlated with poorer survival, while HPV-negative patients had a higher rate of secondary primary tumors (9.12% vs 1.67%)[26].

In a cohort of 46 Egyptian breast cancer patients, targeted next-generation sequencing (NGS) of 409 cancer-related genes revealed TP53 and PIK3CA as the most frequently mutated genes, each found in 58.7% of samples with 15 and 8 distinct somatic mutations, respectively[27]. ClinVar analysis identified 19 pathogenic gene mutations, including 7 in TP53, 5 in PIK3CA, and single pathogenic variants in VHL, STK11, protein kinase B, KRAS, isocitrate dehydrogenase 2 (IDH2), phosphatase and tensin homolog, and ERBB2 (Figure 3). Additionally, they reported 5 variants of uncertain significance (e.g., 4 in TP53, 1 in CEBPA) and 4 with conflicting interpretations (including adenomatous polyposis coli and JAK3), along with a notable drug-response variant, TP53-p.P72R found in 8 cases. Novel variants were also discovered in JAK2, mTOR, KIT, and EPHB, and pathway analysis highlighted phosphoinositol-3-kinase/protein kinase B signaling as altered in over 50% of cases, suggesting it as a potential therapeutic target.

Figure 3 Comprehensive and visually structured overview of the evolution of precision medicine, tracing its development from population-based mutation profiling toward fully personalized, spatiotemporal cancer care. It begins with the paradigm shift from generalized population-level strategies to personalized medicine, where the integration of individual genomic and lifestyle data takes center stage. The next major advancement depicted is next-generation sequencing, enabling broad genomic profiling and the identification of shared mutations across populations, a foundational step in large-scale precision oncology. This leads into the application of single-cell genomics, which addresses cellular heterogeneity by mapping the transcriptional landscapes of individual tumor cells, uncovering subpopulations with distinct biological characteristics. Further refinement is achieved through spatial omics and proteomics technologies, which provide tissue-contextual protein mapping at high resolution, crucial for pathology insights and tumor microenvironment analysis. The figure then emphasizes the role of artificial intelligence-assisted imaging and biomarker modeling, which supports real-time, end-to-end response tracking and dynamic patient monitoring. As precision deepens, the N-of-1 treatment design is introduced, highlighting adaptive trial models tailored to individual patients through the integration of genomic data and lifestyle factors. The final component in the figure encapsulates spatiotemporal assessment of prognosis and monitoring, where therapy responses, resistance patterns, recurrence risks, and lifestyle influences are evaluated across time to enable personalized lifetime cancer care. This integrated, stepwise visualization underscores the convergence of omics technologies, artificial intelligence, and individualized data in revolutionizing modern cancer diagnostics and therapeutics. AI: Artificial intelligence; NGS: Next-generation sequencing.

A certain study analyzed surveillance, epidemiology, and end results data to examine HPV positivity in oral cavity and pharynx cancer across different racial and ethnic groups. Findings revealed that American Indian/Alaska native and Asian or Pacific Islander patients, along with Hispanic/Latino individuals, had lower odds of being HPV-positive compared to White patients, while racial disparities in HPV association persisted more significantly in non-OP oral cavity and pharynx cancers[28]. A large cohort study analyzed 4549 United States patients with recurrent or metastatic HNSCC receiving immune checkpoint inhibitors (ICIs), with 3020 (66.4%) White, 260 (5.7%) Black or African American, and 56 (1.2%) Asian patients[29]. Survival was significantly longer in HPV-positive patients (16.6 months vs 8.8 months in HPV-negative cases), and no major survival differences were observed between those who discontinued vs continued ICI at 1 or 2 years, suggesting that stopping ICI in long-term responders may be a viable option. That study analyzed 9887 OSCC cases in Queensland, Australia, revealing a 4.49-fold increase in diagnoses and a 19.14-fold rise in deaths over 36 years, with indigenous status, low socioeconomic status, and rural residency identified as key risk factors[30]. Males were diagnosed at younger ages, tongue SCC was the most common (49%), and poorer differentiation correlated with worse survival, highlighting the urgent need for targeted screening and early intervention strategies in high-risk Australian populations.

QUESTIONS REMAIN BEYOND PRECISION MEDICINE VS PERSONALIZED MEDICINE TOWARD A MULTIDISCIPLINARY APPROACH TO SINGLE-CELL PROFILING

This article challenges the scientific community to rethink long-standing paradigms. Readers are urged to examine the findings critically, question established connections, and reflect on the broader implications of marker interactions in SCC. Does the contrasting behavior of p16 and p53 suggest a need for cancer-specific biomarker frameworks? How do environmental factors and genetic predispositions reshape these molecular landscapes of individuals?

Personalized medicine vs precision medicine: Clarifying the distinction

Personalized medicine refers to the tailoring of medical treatment to the individual characteristics of each patient. This approach takes into account genetic, environmental, and lifestyle factors to optimize disease prevention, diagnosis, and treatment for a specific person. It contrasts with the traditional “one-size-fits-all” model of medicine by recognizing that individuals respond differently to the same medical interventions[31]. Precision medicine, while often used interchangeably with personalized medicine, is a broader concept that focuses on identifying subpopulations that may respond similarly to a particular treatment. Precision medicine does not necessarily imply an individualized treatment plan but rather groups patients based on shared molecular, genetic, or other biological markers (National Research Council, 2011)[32]. The term “precision medicine” emerged to emphasize reproducibility and data-driven approaches in medicine, but, as critics argue, it lacks the direct patient-centric focus that “personalized medicine” conveys[33]. In summary, while precision medicine aims for accuracy in disease classification and targeted therapies, personalized medicine emphasizes bespoke treatment for an individual patient. The latter term may better capture the essence of truly individualized healthcare interventions (Figure 3) in the technology and concept evaluation. The distinction between personalized medicine and precision medicine is crucial in advancing modern healthcare. While precision medicine leverages molecular diagnostics to identify targetable genetic mutations, personalized medicine provides a more comprehensive, patient-centric framework that accounts for genetic, physiological, environmental, and lifestyle variables. A truly individualized treatment paradigm must integrate both approaches, ensuring that therapies are not only genetically precise but also holistically tailored to each patient’s unique pathophysiology and treatment response variability.

Personalized medicine vs precision medicine: A nuanced distinction in oncology and beyond

In contemporary biomedical research and clinical practice, personalized medicine and precision medicine are often used interchangeably, yet they encapsulate distinct paradigms of patient care. While both approaches aim to tailor treatments to individual patients, their scope and methodology differ in fundamental ways. Personalized medicine emphasizes a holistic view of patient care, incorporating genetic, environmental, lifestyle, physiological, and pathological factors to craft individualized treatment strategies. In contrast, precision medicine primarily focuses on genomic and molecular profiling to identify targetable genetic mutations and biomarkers, often guiding therapy selection based on these specific molecular features.

Beyond genetic mutations: The need for a holistic approach

A key limitation of precision medicine is its reliance on genetic alterations as primary determinants of treatment response. While targeted therapies - such as those inhibiting oncogenic driver mutations [e.g., epidermal growth factor receptor (EGFR) in lung cancer or BRAFV600E in melanoma] - have revolutionized cancer treatment, clinical outcomes among patients harboring identical mutations can vary widely. This discrepancy underscores the critical role of personalized medicine, which accounts for tumor microenvironment interactions, immune response variability, pharmacogenomics, metabolic profiles, comorbidities, and individual physiological conditions that influence therapeutic efficacy. For instance, studies have demonstrated that despite the presence of the same EGFR mutations, patients with non-small cell lung cancer exhibit heterogeneity in their responses to EGFR tyrosine kinase inhibitors[34]. Differences in tumor heterogeneity, immune evasion mechanisms, and metabolic adaptations can alter drug sensitivity, leading to varying degrees of therapeutic benefit or resistance. For example, lung cancer remains the leading cause of cancer-related deaths globally, with 1.8 million new cases and 1.6 million deaths annually[35], while 5-year survival rates range from 4% to 17% depending on stage and region. Here are the problems. While precision medicine has revolutionized lung cancer treatment through targeted therapies, its implementation remains incomplete and insufficient for many patient subgroups. Current approaches primarily focus on molecularly defined oncogenic drivers, such as EGFR mutations or anaplastic lymphoma kinase rearrangements, guiding the use of tyrosine kinase inhibitors[36]. However, this strategy overlooks patient-specific pathological and physiological conditions, leading to inconsistent clinical outcomes among individuals with identical genetic mutations.

Moreover, many lung cancer patients, including those with poor performance status or elderly individuals, are excluded from clinical trials, leaving them without a clear, precision-based treatment framework. This limitation highlights the gap between theoretical precision medicine and its real-world applicability, as these patients often experience treatment-related toxicity or reduced drug efficacy due to underlying comorbidities, immune status, or metabolic differences - factors largely ignored in current precision paradigms. These manifest in the need for a broader, truly personalized precision medicine. True precision medicine should go beyond genomic profiling and integrate tumor microenvironment dynamics, immune system interactions, pharmacogenomics, and individual metabolic responses to optimize therapeutic interventions. For instance, patients with the same driver mutation may exhibit vastly different responses due to tumor heterogeneity, immune checkpoint expression, or gut microbiome composition, all of which influence drug metabolism and resistance mechanisms.

Additionally, while current precision medicine emphasizes late-stage treatment customization, a more effective approach would prioritize prevention and early detection, particularly through lung cancer screening and risk stratification based on comprehensive biomarkers, not just smoking history. Addressing the oversight in smoking prevention and cessation manifested beyond biomarker indications. While smoking remains the leading risk factor for lung cancer, precision medicine should not solely be reactive - it must integrate preventative strategies, including screening and individualized risk assessments. Despite the acknowledgment of smoking prevention’s importance, its exclusion from this discussion underscores the failure of existing precision medicine to adopt a proactive, patient-centric approach that considers behavioral and environmental risk factors alongside genetic markers.

Similarly, even in hematologic malignancies, patients with identical driver mutations may respond differently to targeted treatments due to variations in clonal architecture and microenvironmental influences[37]. Clonal heterogeneity significantly influences disease behavior and drug response in acute myeloid leukemia (AML), as shown in this cohort of 2829 patients, where high variant allele frequency in mutations like NRAS and TET2 correlates with poor prognosis, while elevated GATA2 variant allele frequency is linked to better outcomes. Clinical factors such as white blood cell count and blast percentage are associated with the subclonal abundance of mutations like TP53 and IDH1, and patients with specific mutation order patterns (e.g., cohesin before NPM1) exhibit shorter survival. Interestingly, a branched clonal evolution pattern is linked to improved prognosis, and certain mutations (NRAS, IDH1) predict drug sensitivity, highlighting the need for assessing clonal architecture to refine prognosis and guide therapy selection in AML. The clonal architecture is initiated and constructed in the stem cell niche of bone marrow.

How does the niche anchoring of leukemia stem cells (LSCs) influence AML progression? Genetically deleting junctional adhesion molecule (JAM)-C in a mouse model shows the loss of JAM3 shifted hematopoietic stem cell dynamics from long-term to short-term expansion, altering gene expression without impacting leukemia initiation or progression[38]. RNA sequencing revealed that activating protein-1 and tumour necrosis factor alpha/nuclear factor-kappa B pathways were upregulated in JAM3-deficient leukemic cells, leading to the identification of a novel prognostic gene signature distinct from the LSC-17 score. Patient stratification using both LSC-17 and AP-1/TNF-α gene signatures defined four risk groups with survival rates ranging from less than one year to over eight years, emphasizing the role of LSC anchoring in prognosis. These findings suggest that bone marrow niche interactions actively rewire leukemic gene programs, improving risk assessment and potential therapeutic targeting in AML. However, little is known about JAM3-deficient leukemia in patients.

Personalized pathological and physiological conditions: Key to individualized treatment

Beyond genomic data, individualized physiological conditions significantly influence treatment efficacy and patient outcomes. Factors such as chronic inflammation, microbiome composition, immunocompetency, and metabolic status can dictate how a patient metabolizes and responds to a given therapy. For example, research has highlighted the impact of gut microbiota on ICI efficacy, with specific microbial compositions correlating with enhanced or diminished responses in cancer immunotherapy - influencing the efficacy of immune-checkpoint inhibitors [cytotoxic T-lymphocyte antigen-4, programmed death-1 (PD-1)/programmed death ligand-1] and immunogenic chemotherapies[39]. The specific microbial compositions, including Akkermansia muciniphila, Bacteroides fragilis, Bifidobacterium spp., and Faecalibacterium spp., are linked to enhanced immune surveillance and favorable cancer treatment outcomes in both preclinical and clinical settings. Beyond oncology, a diverse and balanced microbiome supports overall health and immune homeostasis, reducing the risk of metabolic and inflammatory disorders, underscoring its broader therapeutic significance. This finding suggests that while precision medicine may identify genetic markers predicting response, personalized medicine integrates host-related factors to refine and optimize treatment decisions.

Similarly, the personalized response to drug therapy is increasingly driven by cytochrome P450 pharmacogenetics, which has identified key genetic variants influencing individual drug metabolism and adverse reactions. This progress enables predictive pharmacogenetics, allowing for dose adjustments and drug selection tailored to a patient’s genetic makeup, thereby improving both efficacy and safety. Pharmacogenomic studies further reveal that drug metabolism varies widely among individuals due to polymorphisms in genes encoding metabolic enzymes such as CYP2D6 and CYP3A4, which significantly impact drug bioavailability and toxicity[40]. This variability underscores the limitations of a one-size-fits-all approach, highlighting the need for precise, individualized dosing strategies rather than relying solely on genetic mutations. As genotyping becomes more accessible, P450-based personalized treatments could transform clinical practice, optimizing 10%-20% of all prescriptions, reducing costs, and significantly improving health outcomes through precision medicine. In a phase II trial of 113 patients with hormone-receptor-positive metastatic breast cancer treated with tamoxifen, no significant association was found between CYP2D6 metabolizer status or circulating endoxifen levels and progression-free survival (PFS)[41]. Among 85 evaluable patients, poor metabolizers had a median PFS of 12.9 months compared to 6.9 months in intermediate/normal metabolizers, while patients with high vs low-endoxifen concentrations had median PFS of 13.8 months vs 11.1 months, though the small sample size limited definitive conclusions.

Bridging the gap: Integrating personalized and precision medicine

A synergistic model incorporating both precision and personalized medicine offers the most comprehensive strategy for individualized patient care. While precision medicine identifies molecular vulnerabilities for targeted intervention, personalized medicine ensures these interventions are tailored to the patient’s broader physiological and pathological context, e.g., neuroblastoma’s heterogeneity limits the predictive power of current Children’s Oncology Group risk stratification, often leading to overtreatment. By integrating an intra-tumor microbial gene abundance score (M-score), researchers identified two subgroups within high-risk patients - M (high) and M (low) - revealing distinct survival outcomes and implicating cyclic adenosine monophosphate response element binding protein over-activation in tumor progression and metastasis[42]. Future advancements in multi-omics integration, artificial intelligence (AI)-driven predictive modeling, and real-time biomarker monitoring will further enhance this hybrid approach, optimizing therapeutic outcomes at the individual level.

Unraveling individual cellular identities through single-cell and gene expression genomics elucidates how cellular heterogeneity and transcriptional landscapes shape personalized disease trajectories and therapeutic vulnerabilities

The scRNA-seq has significantly advanced our understanding of intratumoral heterogeneity by identifying tumor subpopulations with distinct biological characteristics, thereby informing more precise therapeutic strategies, including that “scRNA-seq highlights intratumoral heterogeneity in primary glioblastoma”[43]; that “Intratumor heterogeneity of EGFR expression mediates targeted therapy resistance and formation of drug tolerant microenvironment”[34]; that “Single-cell RNA landscape of intratumoral heterogeneity and immunosuppressive microenvironment in advanced osteosarcoma”[44]; that “scRNA-seq reveals distinct patterns of cell state heterogeneity in mouse models of breast cancer”[45]; and that “Additionally, scRNA-seq has revealed precise transcriptional changes during neuroinflammation at the single-cell level, opening new avenues for exploring disease mechanisms and drug discovery in conditions such as stroke”[46]. Glioblastoma (GBM) multiforme exhibits subclonal evolution, contributing to its aggressiveness and resistance to therapy. By analyzing single-cell transcriptomes from primary and relapsed tumors, researchers identified mutations in the RAS/guanine-nucleotide-exchange factor (GEF) signaling pathway - validated further by meta-analysis of RNA-seq data from more than 3000 + patients - highlighting its role in GBM relapse and its potential as a target for precision therapy[47]. Notably, only single-cell molecular analysis overcomes the inherent heterogeneity of conventionally used bulk tumors with respect to defining tumor subclonal evolution relevant to GBM recurrence.

Exploring high-resolution protein mapping

High-resolution protein mapping within tissue contexts enhances our understanding of patient-specific pathology and propels precision diagnostics forward. For instance, a study employing spatially resolved proteomics achieved deep coverage by quantifying over 9000 proteins at approximately 500 μm resolution in mouse brain tissues, offering insights into spatial protein heterogeneity crucial for understanding neurodegenerative diseases[48]. Additionally, the Human Protein Atlas (HPA) project integrates various omics technologies to map protein expression across human tissues, providing a valuable resource for exploring tissue-specific protein distribution and its implications for health and disease (https://www.proteinatlas.org/ - the HPA is a Swedish-based program started in 2003, and the HPA consortium is funded by the Knut and Alice Wallenberg foundation, accessed on March 21, 2025).

Reflecting on the transformative acceleration of NGS, spatialomics, and single-cell biology - and their collective potential to redefine personalized genomic medicine, shifting the paradigm from population averages to individual molecular fingerprints, which evolve with lifestyle and time. The rapid advancement of NGS, spatial omics, and single-cell biology is revolutionizing personalized genomic medicine by enabling a shift from population-based approaches to individualized molecular profiling[49]. NGS technologies have significantly enhanced our understanding of genetic and pathogenesis factors in diseases like cancer, facilitating the development of precision medicine strategies tailored to individual genetic profiles. Concurrently, spatial omics technologies have emerged, allowing for the comprehensive analysis of molecular characteristics within their native tissue contexts and highlighting the integration of spatial omics with high-throughput sequencing, reinforcing the shift from generalized population models to personalized molecular profiles that evolve with tissue context and temporal dynamics, a cornerstone of precision medicine, thereby refining our understanding of cellular heterogeneity and tissue organization[50]. These advancements collectively contribute to a more precise and dynamic approach to medical treatment, moving beyond generalized population averages to address the unique molecular fingerprints of individuals.

The emerging N-of-1 paradigm in radiation oncology reflects a shift from standardized protocols to highly individualized treatment plans tailored to each patient’s tumor biology, anatomy, and response. Innovations such as surface-guided radiotherapy improve setup accuracy and patient comfort, enabling more precise delivery with less variability[51]. Proton beam and magnetic resonance guided radiotherapy allow for real-time adaptation and tissue-sparing dose escalation, which can dramatically impact survival as simulated with an AI model[52], as seen in intrahepatic cholangiocarcinoma cases[53]. The integration of AI and advanced imaging supports dynamic, patient-specific treatment adjustments in cancers like nasopharyngeal carcinoma. A separate study on nasopharyngeal cancer documents the integration of personalized treatment plans with advanced imaging and AI for better treatment response assessment[54]. These advances, combined with the basic mouse models, such as that Piezo1-mediated mechanosensation in bone marrow macrophages promotes vascular niche regeneration after irradiation injury[55] and that CRIF1 is a potential target to improve the radiosensitivity of osteosarcoma[56], all collectively underscore a move toward single-patient-oriented radiotherapy, where personalization isn’t an exception, but a new clinical expectation.

Personalized biomarker signature monitoring cancer subclonal evolution

In a study involving 260 blood samples - 100 from an internal diffuse large B-cell lymphoma cohort and 160 from an external validation cohort - a 6-tRNA-derived small RNA classifier was developed and validated using reverse transcription real-time PCR[57]. The classifier demonstrated strong prognostic power, distinguishing high-risk from low-risk patients [hazard ratio = 6.657, 95% confidence interval (CI): 2.827-15.68, P = 0.0006] and accurately differentiating diffuse large B-cell lymphoma patients from healthy individuals (area under the receiver operating characteristic curve = 0.882, 95%CI: 0.826-0.939). Notably, a dynamic analysis revealed that patients with a ≥ 1.06-fold decrease in the classifier score after one therapy cycle had improved treatment outcomes.

THE N-OF-1 PARADIGM BEYOND RARE GENETIC DISORDERS: A PATIENT-CENTRIC FUTURESPATIALLY RESOLVED PROTEOMIC PROFILING COMPOSED OF SPATIALLY ORGANIZED, HETEROGENEOUSLY DISTRIBUTED CELLS THAT INTERACT WITH EACH OTHER AND THE SURROUNDING EXTRACELLULAR MATRIX

The N-of-1 paradigm represents a transformative shift in personalized medicine-moving beyond population-based averages to embrace the unique biological, clinical, and lifestyle profiles of individual patients. Initially championed in rare genetic disorders, this approach is gaining broader relevance across chronic conditions, neurodevelopmental disorders, and oncology. By integrating patient-specific outcome data with adaptive, trial-like frameworks, N-of-1 treatment plans enable clinicians to make more precise, evidence-guided therapeutic decisions. Clinical studies demonstrate the power of this model: In attention-deficit hyperactivity disorder and chronic disease, up to 54% of treatment decisions aligned with N-of-1 trial outcomes[58]; in oncology, personalized molecular profiling led to longer PFS - the use of individualized molecular profiling within tumor boards led to longer PFS and overall survival (OS), as described by Kato et al[59] and Sicklick et al[60] for molecular profiling of cancer patients enables personalized combination therapy: The I-PREDICT study[60]; and in neurodegenerative conditions, individualized interventions yielded measurable symptom relief - in rare conditions, single-patient adaptations yielded symptom reductions, such as seizure decreases in a neurodegenerative disorder[61]. Despite methodological complexities, advances in adaptive trial design, real-world data integration, and statistical innovation are making N-of-1 not only feasible, but essential - charting a new course toward truly individualized care for every patient, not just the rare few.

To enhance precision medicine, early integration of molecular matching and comprehensive multi-omic profiling is crucial. Recent advances in biotechnology have unveiled complex molecular features driving malignancies, revolutionizing precision medicine. By leveraging NGS and targeted therapies, the outlook for several fatal cancers has dramatically improved. Tumor and liquid biopsy profiling, along with transcriptomic, immunomic, and proteomic analyses, now guide optimized treatments. New trial designs, such as basket, umbrella, master platform, and N-of-1 patient-centric studies, are replacing traditional phase I-III protocols, facilitating faster drug evaluation and approval. Real-world data, digital tools, and machine learning/AI further accelerate knowledge acquisition. Clinical trials have shifted from tumor type-based to gene-directed, histology-agnostic approaches, with personalized combination therapies tailored to individual biomarker profiles. Some novel trials show that matched therapies outperform non-matched treatments across various cancers[62].

N-of-1 trials offer an exceptional opportunity for fast-tracking of cancer drug development and treatment[63]. Personalized medicine, including lifestyle interventions like diet and exercise, can significantly impact cancer treatment by tailoring therapies to the individual. N-of-1 trials, which focus on single-patient data, help resolve therapeutic uncertainties, potentially improving patient outcomes compared to standard care. This approach is especially valuable for ultra-rare or nano-rare diseases lacking established therapies. However, challenges exist in the regulatory framework for personalized RNA therapies, as these approaches may not yet be universally applicable or well-supported by traditional clinical trial methods. The transformative potential of personalized treatments is clear, but scalability and regulatory hurdles remain.

Unveiling the prognostic complexity of p16, p53, and HPV in HNCSCC demands that an ethical integration is essential for a paradigm shift from standardized to individualized N-of-1 treatment plan can be initiated by physicians. Our recent physician-guided study shows that triple biomarkers manifested in one patient screened out of 26 patients, which implies that the treatment resistance developed only in this patient[64]. Chromosomal rearrangements drive leukemogenesis in 50% of AML cases, yet targeted therapies remain limited due to a lack of reliable biomarkers for treatment resistance and relapse. This study analyzed 16 patients with chromosomal rearrangements and identified a previously unknown CCDC32/CBX3 gene fusion in KMT2A/AFDN-rearranged AML, which persisted in a 21-year-old patient with rapid relapse, influencing wild-type gene expression and promoting treatment resistance. Notably, while some patients lost the KMT2A/AFDN fusion upon relapse, the CCDC32/CBX3 fusion persisted, suggesting its role in subclonal evolution and therapy resistance. Functional validation in MV4-11 AML cells confirmed that CCDC32/CBX3 overexpression drives cell cycle progression, providing experimental evidence of its oncogenic potential. This study identifies a triple biomarker signature - KRAS mutation, dual gene fusions (KMT2A/AFDN, CCDC32/CBX3), and chimeric RNA variants - which may serve as a prognostic and predictive tool for poor clinical outcomes in AML. Thus, precision medicine gravitates toward an individualized N-of-1 treatment platform.

Not only focusing on an individual but also focusing on a specific subclone of the individual patient - this work connects to the evolution of cancer care, where clinicians can now analyze specific biological markers to tailor treatment plans for individual patients’ specific single cells (clone and subclone)[16]. In one study, researchers found that certain levels of specific subclonal activation predicted a positive response to targeted therapy. Identifying these patterns early can help avoid the traditional trial-and-error approach of switching drugs, which often leads to patient frustration and discontinuation of care. This is the kind of research that has the potential to change lives. Personalized treatment will be the cornerstone of 21^st-century cancer care, guiding a more precise and proactive approach to therapy. For instance, Li and Kabeer[65] explored the concept of spatiotemporal switching signals in cancer stem cell activation, particularly in pediatric-origin cancers that manifest in adulthood. Their findings suggest a “watch-and-wait” lifetime strategy for cancer treatment, emphasizing the need for long-term, personalized monitoring rather than reactive interventions. Such subclonal evolution also manifested in HNCSCC[15].

Another study supports the shift toward personalized oncology by demonstrating how single-cell molecular profiling of circulating tumor cells (CTCs) can identify patient-specific dysregulated pathways in HNSCC[66]. Such insights hold the promise of improving early detection, tailoring targeted therapies, and monitoring treatment response - all fundamental principles of precision medicine in cancer care. More specific highlights show the potential of personalized medicine in treating HNSCC by leveraging single-cell molecular profiling to identify dysregulated signaling pathways associated with CTCs[67]. Personalized medicine aims to tailor treatments based on an individual’s unique molecular and genetic profile, and this research aligns with that goal in several key ways: (1) Molecular characterization for targeted therapies: By identifying the gene expression and mutational landscapes of CTCs using scRNA-seq, the study provides patient-specific insights into tumor progression and metastasis. This information can guide the development of precision therapies that target these pathways, reducing the reliance on broad-spectrum treatments; (2) CTCs as personalized biomarkers: The study underscores the role of CTCs as real-time liquid biopsy markers, allowing for non-invasive monitoring of disease progression. Personalized medicine benefits from such approaches by enabling dynamic treatment adjustments based on the evolving molecular profile of an individual’s cancer; (3) Targeting treatment-resistant cells: One of the biggest challenges in HNSCC is treatment resistance, which leads to poor survival rates. The identified five dysregulated signaling pathways could serve as new therapeutic targets, enabling the design of customized regimens that address an individual patient’s resistance mechanisms; (4) Screening and early detection for personalized risk assessment: The study’s findings may lead to the development of personalized biomarker panels for early HNSCC detection, allowing for risk stratification and tailored interventions before metastasis occurs; and (5) Integration with existing cancer genomic data with new technologies: By cross-validating findings with previously published studies, this research strengthens the foundation for integrating personalized molecular diagnostics into standard HNSCC treatment protocols, paving the way for precision oncology approaches.

Along the same vein, a recent study introduces scIDUC, a computational framework that integrates single-cell transcriptomic data with pan-cancer drug screening datasets to predict therapeutic vulnerabilities at the individual-cell level[68]. It demonstrated high accuracy (Cohen’s d > 1.0) in stratifying drug sensitivity, enabling identification of resistant subpopulations across tumor types like rhabdomyosarcoma, pancreatic ductal adenocarcinoma, and castration-resistant prostate cancer. scIDUC successfully pinpointed alternative drugs for cell subsets unresponsive to standard-of-care treatments, including resistance arising from intrinsic tumor biology or microenvironmental influences. The predictions were validated experimentally, showcasing scIDUC’s utility in tailoring treatments based on tumor heterogeneity and guiding precision oncology (Figure 3).

Mammalian organs and tissues are composed of spatially organized, heterogeneously distributed cells that interact with each other and the surrounding extracellular matrix (ECM). As proteins mediate nearly all cellular functions and serve as key biomarkers and therapeutic targets, spatially resolved proteomic profiling is essential for understanding cell phenotypes and disease mechanisms beyond genomic and transcriptomic alterations. The study introduces S4P, a sparse sampling strategy powered by deep learning to enable high-throughput spatial proteomics across whole tissue slices, overcoming traditional mass spectrometry limitations[68]. Using this method, researchers mapped over 9000 proteins in the mouse brain, generating the most comprehensive spatial proteome to date. The approach enables the detection of regional and cell type-specific protein markers, addressing challenges in identifying functional phenotypes driven by gene mutations and transcriptome alterations. Given its improved sensitivity and efficiency, S4P holds promise for spatial proteomic profiling across diverse tissue types and disease contexts.

Integrated AI algorithm highlights how spatiotemporal omics technologies are transforming our understanding of how genetic information is translated into dynamic, heterogeneous cellular systems across space and time. By combining spatial transcriptomics, proteomics, and imaging-based omics, researchers can now map the distribution, interaction, and evolution of specific cell types, gene expression patterns, and biomarkers within tissues, uncovering mechanisms of development, disease, and regeneration[69]. Large-scale studies such as genome-wide association studies and the 1000 genomes project have identified numerous variants, but spatial omics adds critical resolution by showing how these variants function at the cellular and organ level, offering insights into diseases beyond bulk-genome association (Figure 3). This platform also points to emerging AI subtypes, such as deep learning-enhanced tissue reconstruction and predictive modeling, as key enablers of high-resolution data integration and functional interpretation. Together, spatiotemporal omics and AI promise a paradigm shift toward ecosystem-level biology, enabling precise, location- and time-specific biomarker discovery and therapeutic targeting (Figure 3).

CANCER AND HOST: FROM PARASITISM TO CO-EVOLUTION OF MUTUALISM

“Nothing in biology makes sense except in the light of evolution”, with this iconic declaration, Dobzhansky[70] titled his seminal 1973 essay, positioning evolution as the unifying principle of the life sciences. Dobzhansky[70] argued that without the framework of evolutionary theory, biological observations would remain a disjointed collection of unrelated facts, lacking coherence or explanatory power. Evolution, in his view, is the central pillar that transforms the diversity of life, the unity of organisms, and the complexity of biological phenomena into an intelligible whole. The essay was also a response to anti-evolutionary and creationist arguments, defending the indispensability of evolution in education and research. By demonstrating how concepts such as common ancestry, homologous structures across species, and the fossil record all become comprehensible under the lens of evolution, Anderson et al[71] showed that this theory provides not only explanatory breadth but also predictive strength. Borrowing the phrase “light of evolution” from Jesuit priest and paleontologist Pierre Teilhard de Chardin, Anderson et al[71] elevated the idea into a guiding maxim for modern biology. The essay has since become a classic, widely cited as essential reading in evolutionary and population biology, reminding generations of scientists and students alike that evolution is not just one branch of biology - it is the discipline’s unifying framework.

Shedding the “light of evolution” on cancer and host: From parasitism to co-evolution of mutualism, leads us to rethink - for decades, cancer has been cast as a fatal, destructive parasite - an enemy that deprives the host without restraint - a rogue growth that consumes its host until death. But this view is too narrow. Emerging evidence and conceptual frameworks suggest that cancer is not only an adversary but also a part of a coevolutionary dialogue with the host body - part of a continuum of host-organism relationships, spanning from parasitism to mutualism (Figure 4). Thus, when viewed through the broader lens of biology, the cancer conceptualization shift opens new conceptual framework pathways for understanding its role in evolution, one that can illuminate new strategies for therapy.

Figure 4 Spectrum of host-organism relationships, applied to cancer. The continuum of biological interactions ranges from mutualism (benefiting both partners; e.g., gut microbiota), through commensalism (benefiting one partner without harming the other; e.g., skin flora), to parasitism (benefiting one at the expense of the other; e.g., malaria parasite). Early cancer is positioned closer to parasitism but not at the extreme, reflecting its dual role as both a host threat and a potential trainer of immunity. Two conceptual frameworks are overlaid: (1) Cancer niche as a “garbage disposal machine”, Li et al[72], 2019, for subclones, Li et al[73], 2012; green annotation: Highlighting how tumors may buffer metabolic imbalance and function transiently as regulatory sinks within the body-disease continuum, echoing traditional Chinese medicine concepts of balance; and (2) Subclonal switchboard signaling, Li et al[73], 2012; blue annotation: Emphasizing that dominant and dormant cancer subclones dynamically exchange roles, shifting the tumor’s position along the spectrum and driving progression or recurrence. Together, these perspectives reframe cancer not merely as a destructive parasite, but as part of a coevolutionary dialogue with the host, spanning the continuum of mutualism, commensalism, and parasitism.

Lessons from mutualism in humans: Cancer as an early alarm and trainer

In ecology, mutualism describes interactions where both species benefit. Humans are deeply embedded in such relationships: (1) Gut microbiota provide essential functions in digestion, metabolism, and immunity, while humans supply nutrients and habitat; and (2) Even head lice, though parasitic, may once have played an indirect protective role - eliciting immune responses that reduced vulnerability to deadly diseases transmitted by body lice. These examples reveal a spectrum: Some organisms exploit us, others sustain us, and sometimes the boundary blurs. What appears harmful in one context may be adaptive in another. Cancer may work as an early alarm and trainer - a single transformed cell is not merely the seed of destruction. By displaying aberrant antigens and disrupting tissue homeostasis, transformed cells act as an alarm signal, activating immune surveillance - strengthening surveillance and resilience. These earliest encounters function like training drills, forcing the immune system to sharpen its recognition and response. Thus, cancer at its inception can paradoxically serve as a survival exercise for the host, preparing immunity for future threats.

Cancer as a co-evolutionary partner

Likewise, cancer may not be only a parasite at its earliest stages, like a niche as a garbage disposal: Tumors can function as buffers, absorbing metabolic and cellular imbalance much like commensal partners absorb toxins in ecosystems[72]. In this light, the tumor microenvironment may function as a garbage disposal machine, a sink for metabolic waste, senescent cells, and dysregulated signals. Far from being purely destructive, the niche can temporarily absorb imbalance, maintaining a fragile equilibrium between body and disease. Traditional Chinese medicine frames this as restoring harmony - not erasing all challenge but rebalancing the continuum of health and illness. Equilibrium and coexistence: Host and tumor can enter temporary stalemates, echoing the balance seen in stable host-microbe relationships. “Why This Matters?” This conceptual shift in perspective matters because it changes cancer from a one-way march to destruction into a reciprocal process. Early cancers strengthen immunity; tumor niches buffer imbalance; the host adapts under selective pressure. Cancer is dangerous, but its earliest stages may confer resilience and understanding that resilience could inspire new therapeutic designs.

Subclonal dynamics and host adaptation

As illustrated by the concept of subclonal switchboard signaling, cancer progression is a dynamic process[73]. Destroying one dominant subclone can awaken dormant ones, much as ecological niches shift when one species is removed. This constant back-and-forth mirrors the co-evolutionary pressures in symbiosis - cancer adapts, the host counters, and the struggle for dynamic balance vs deadlock shapes both. Western oncology has often pursued an annihilation strategy, attempting to eradicate tumors completely. Yet, decades of focusing on bulk heterogeneity have brought limited progress, particularly in solid tumors like GBM (refer to “Standard therapies and response percentage rates of population, an argument for N-of-1 medicine to balance precision medicine and personalized medicine, gravitates toward N-of-1 scenario medicine”)[47]. By reframing cancer as part of a body-disease continuum, we see that host and tumor exist in a dynamic equilibrium, sometimes coexisting in a state of stalemate, sometimes tipping toward progression.

Mutualism, parasitism, and the continuum of life: In nature, the line between parasite and mutualist is not fixed

Microbes in our gut, once pathogens, may evolve into indispensable allies. Lice, though harmful in isolation, may have provided protection against worse diseases. Cancer, while devastating at advanced stages, may play a role in immune education, waste clearance, and systemic balance at earlier stages. Cancer, like microbes and parasites, belongs on the spectrum of host relationships - from parasitism to mutualism (Figure 4). In its earliest forms, it may train immunity, buffer imbalance, and force adaptation, echoing the way gut flora sustains us or even lice once shielded us. At its worst, it becomes a runaway parasite. Understanding cancer through this ecological lens - as part of a coevolutionary continuum - can inspire therapies aimed not only at eradication but at restoring balance and harnessing the adaptive lessons cancer offers. Subclonal switchboard signaling and recurrence[73]: At the heart of recurrence lies the subclonal switchboard. Tumors are ecosystems of dominant and dormant subclones. Destroying one dominant clone can inadvertently activate dormant subclones, fueling progression and relapse. We postulated that tumors communicate through “switchboard signals” that orchestrate these shifts, much like stem cell populations balancing quiescence and differentiation. Eliminating or disrupting these signals - rather than indiscriminately targeting all subclones - offers a strategy to prevent recurrence. We can investigate LOCKTACs/“sustained proximity” paradigm. “Load and lock: An emerging class of therapeutics that influence macromolecular dissociation”[74] and develop likewise “load-and-lock” therapeutics that slow the right dissociations at the right interfaces - turning kinetic “speed bumps” into barriers that stall clonal handoffs, harden dormancy, and block escape. We propose fusing LOCKTACs with our subclonal switchboard model, framed as “load-and-lock” therapeutics to influence cancer-host dissociation, plus concrete, testable barrier models.

Load-and-lock: Therapeutics that tune cancer-host dissociation

What we borrow from LOCKTACs: Deshaies and Potts[74] highlight a class of agents that reduce the dissociation rate (koff) of native macromolecular complexes - “LOCKTACs” - to extend complex lifetime and thereby amplify or inhibit pathway output. Unlike classic inhibitors (often blocking kon), LOCKTACs stabilize pre-existing, functionally relevant contacts; context determines whether the effect is agonism or antagonism[74]. Examples span splicing activation [e.g., risdiplam stabilizing U1 snRNP - survival motor neuron 2 (SMN2) pre-mRNA] and motor inhibition [e.g., sovilnesib “freezing” kinesin family member 18A (KIF18A)’s bind-and-release cycle], including key import: Pharmacology by sustained proximity - locking dynamic complexes long enough to redirect biology.

What we add from our subclonal switchboard: Our model posits that tumors are ecosystems of dominant and dormant subclones, whose “switchboard signals” coordinate clonal handoffs; killing one dominant clone can unlock a dormant successor and drive recurrence. The therapeutic pivot is not “kill everything”, but JAM the switchboard - disrupt or hold specific handoffs so dormant clones stay dormant[73].

The load-and-lock framework: (1) Load = pre-position a targeted stabilizer (antibody, glue, bifunctional, peptide, scaffold) at a specific interface that governs a malignant “handoff”; (2) Lock = reduce koff at that interface to stall the subclonal, microenvironmental, or migratory transition that fuels progression/escape; and (3) Goal = create load-and-lock barriers - kinetic “speed bumps” that re-time or prevent the dissociations tumors need to evolve, invade, or recur.

Five speculative load-and-lock barriers (testable): To conceptualize a novel framework for stalling tumor progression, the “five speculative load-and-lock barriers”[9] model proposes a set of mechanistically distinct yet testable interventions aimed at kinetically trapping tumor-favorable interactions. At its core, the strategy leverages a “load-and-lock” approach - pre-positioning molecular stabilizers (load) and extending complex lifetimes via reduced dissociation rates (lock) - to disrupt key cellular dynamics that enable cancer growth, metastasis, and immune evasion (Figure 5). The five barrier types - dormancy, immune synapse, motility, signal-noise, and splice-state - each represent unique cellular interfaces or molecular assemblies where locking interactions could tilt the balance away from malignant behaviors. These barriers are designed to selectively impair dynamic switches in tumor cells, offering potential readouts and therapeutic entry points across various tumor contexts. The accompanying figure visually summarizes this integrative concept, highlighting targets, rationales, and molecular checkpoints for each barrier type: (1) Dormancy barrier - lock the “stay-quiet” contacts. What to lock: Tumor-niche interactions known to maintain quiescence (e.g., adhesion/Notch-like or CXCL12-retention axes in quiescent niches). Rationale: Prolonging these contacts keeps dormant subclones dormant, jamming the switchboard handoff. Readouts: Single-cell lineage tracing of dormancy markers (NR2F1, p27), scRNA-seq/spatial, relapse-free survival after cytotoxic pulses. Conceptually, the mechanism will be indication-specific; locking can be achieved with bispecific “clamps” or molecular glues; (2) Immune synapse barrier - lock cytotoxic synapses; unlock checkpoints. What to lock: T cell-tumor immune synapse residency (e.g., LFA-1-ICAM-1 avidity) to extend contact time; conversely, prevent PD-1-programmed death ligand-1 rebinding by locking PD-1 in an off-interface (context-guided)[60]. Rationale: Increase cytolytic dwell time; avoid kinetic “slippage” that lets targets escape. Readouts: Live-cell synapse dwell (total internal reflection fluorescence), Ca²⁺ flux, killing assays; in vivo TIL residency; (3) Motility barrier - lock motors; stall invasion. What to lock: Motor protein KIF18A - microtubule engagement to halt-step-release cycles and trap chromosomal instability (CIN)-high cells in mitosis (sovilnesib/next-gen KIF18A LOCKTACs) - CIN and TP53 mutations are regulated[61]. Rationale: Invasive growth and CIN depend on rapid motor turnover; slowing koff disables the engine. Readouts: Mitotic delay, spindle phenotypes, CIN-selective cytotoxicity, metastasis assays; (4) Signal-noise barrier - lock latent inhibitors; starve growth cues. What to lock: Latent complex pairings (e.g., keep transforming growth factor-β in its lipid accumulation product/ECM cage), or trap growth factors on engineered decoys[74] to reduce effective dissociation into the tumor field. Rationale: Tumoral plasticity feeds on fast on/off signal flux (vascular endothelial growth factor trap[75], locking away ligands drops pathway signal-to-noise ratio[76,77]. Readouts: Ligand-free vs bound pools (bio-layer interferometry/surface plasmon resonance), phospho-signatures, epithelial-mesenchymal transition reversal; and (5) Splice-state barrier - lock tumor into a less-fit isoform state. What to lock: Spliceosome/pre-mRNA assemblies that bias malignant isoform programs toward less invasive states (risdiplam is a clinical precedent for locking splice assemblies)[78]. Rationale: Many oncogenic behaviors are isoform-encoded; stabilizing benign splice complexes hardens the state. Readouts: Isoform ratios (RNA-seq), phenotypic assays (migration, invasion), fitness under therapy pressure. Experimental evidence shows that an SMN2 pre-mRNA splicing modifier, risdiplam, increases the production of full-length SMN protein, the lack of which drives the pathophysiology of spinal muscular atrophy. Risdiplam showed durable (24-month) motor improvement and a generally well-tolerated profile in FIREFISH/SUNFISH clinical trials, offering a convenient oral option across ages and spinal muscular atrophy subtypes[79].

Figure 5 Conceptual diagram of five speculative load-and-lock barriers (testable). This illustration summarizes a unifying load-and-lock strategy designed to kinetically trap tumor-promoting cellular states by reducing the dissociation rate (k_off) of critical molecular interactions. At the center, the strategy emphasizes stabilizing complexes at key interfaces to extend residence time and block rapid transitions. A-E: Radiating from this core are five mechanistically distinct barriers: Dormancy barrier: Locks “stay-quiet” tumor-niche contacts to maintain subclone quiescence and prevent reactivation (A); immune synapse barrier: Locks cytotoxic T cell-tumor synapsed well while blocking immune checkpoint rebinding to enhance killing (B); motility barrier: Locks mitotic motors like KIF18A to stall cell cycle progression and metastatic invasion (C); signal-noise barrier: Locks latent growth inhibitors and decoys ligands to reduce signaling noise and plasticity (D); splice-state barrier: Locks spliceosomal assemblies to bias tumors toward less-fit isoform states (E). Each barrier is represented with a simplified icon and linked to the central load-and-lock concept, illustrating how kinetic stabilization across diverse molecular axes can synergistically constrain tumor evolution. ¹k_off is the “escape rate” of a molecular handshake. Reducing it means tightening the grip, increasing residence time, and functionally locking tumor biology into less dangerous states. In molecular binding kinetics, k_off (the dissociation rate constant) describes how quickly a bound complex falls apart. Mathematically, it has units of per second and represents the probability per unit time that a complex (e.g., protein-ligand, receptor-drug, or synapse contact) will dissociate back into its separate components. A lower k_off = slower unbinding - the interaction is more stable, with longer residence time (the duration the complex stays intact). This is distinct from k_on (association rate constant), which describes how quickly binding occurs. The load-and-lock strategy is essentially about slowing k_off across different molecular systems. By making complexes last longer (kinetically trapping them), you extend functional effects: (1) Dormancy niches hold tumor cells in quiescence; (2) T cell-tumor synapses persist longer, boosting killing; (3) Motor proteins like KIF18A remain stuck, blocking mitosis; (4) Growth inhibitors stay bound, muting noise; and (5) Spliceosomes stay locked in certain states, biasing isoform output.

How “load” meets the switchboard

Subclone-addressable clamps: Bi-functionals that recognize a dominant-clone antigen and lock its required exit interaction (e.g., a cadherin/Notch contact), preventing the dormant clone’s takeover.

Kinetic and gates: LOCKTACs that activate only when two signals coincide (e.g., hypoxia + antigen), reducing off-tumor locking.

Niche pre-loading: Depot or ECM-anchored LOCKTACs that sit in the microenvironment and engage only when a tumor tries to dissociate. Why this should curb recurrence: By holding critical dissociations in place (or locking inhibitory ones), the subclonal switchboard can’t flip to the next dominant program, reducing post-treatment rebound[45].

Assays and development playbook (do-now items)

Kinetics first: Quantify koff shifts with surface plasmon resonance/bio-layer interferometry/stopped-flow on the exact target interfaces highlighted above; prioritize targets where pathway output is koff-sensitive.

Single-cell fate maps: Combine barcoded lineage tracing with spatial transcriptomics to see if load-and-lock halts clonal handoffs in situ.

Function under pressure: Measure dormancy maintenance, immune killing, dwell, CIN-selective mitotic arrest, epithelial-mesenchymal transition reversal as barrier-specific endpoints.

Safety/context: Because locking can agonize or antagonize depending on context, run bidirectional phenotyping (does locking a contact help or harm?) and incorporate reversibility switches (photocleavables, protease-sensitive linkers).

Why does this expand the target space?

LOCKTAC logic targets surfaces we rarely drug (interfaces, multiprotein assemblies), avoids direct competition with native ligands, and operates through time (residence) rather than just occupancy - a complementary axis to proteolysis-targeting chimeras/glues[74]. Integrated insight: Cancer is not merely a parasite. It is both alarm and adversary, garbage disposal and growth, dominant clone and dormant threat. The host body and cancer are locked in a coevolutionary dialogue of challenge, adaptation, and balance. By understanding this dialogue - immune training, niche buffering, equilibrium dynamics, and subclonal switchboard signaling - we can begin to design therapies that move beyond eradication, toward rebalancing and reprogramming the host-cancer relationship.

STANDARD THERAPIES AND RESPONSE PERCENTAGE RATES OF POPULATION, AN ARGUMENT FOR N-OF-1 MEDICINE TO BALANCE PRECISION MEDICINE AND PERSONALIZED MEDICINE, GRAVITATES TOWARD N-OF-1 SCENARIO MEDICINE

A nuanced point about the limitations of population-level statistics in real-time, personal medical decision-making, especially in the face of a life-threatening diagnosis like GBM, provided data with our argument for N-of-1 medicine and the inadequacy of probabilistic communication in the clinic, as shown in “Characterizing benefit from temozolomide in O6-methylguanine-DNA methyltransferase (MGMT) promoter unmethylated and methylated GBM: A systematic review and meta-analysis”[80]. Key points include: (1) This meta-analysis provides pooled survival estimates for GBM based on MGMT status; (2) Pooled median OS: 14.11 months and 24.59 months for unmethylated and methylated GBM, respectively; and (3) Whether temozolomide (TMZ)^® can be omitted in unmethylated GBM cannot be answered definitively.

Population data vs personal decision: The case for N-of-1 medicine

When a patient with GBM asks, “Will this treatment help me?”, the clinician might point to numbers: For example, a median OS of 14.11 months for those with unmethylated MGMT status treated with radiation and TMZ, or 24.59 months for those with methylated MGMT. But to the patient, numbers like 13.18% (the lower bound of the 95%CI for OS) are meaningless in isolation. That number does not answer what the patient truly wants to know: Yes or no - do I have a chance? This is the critical flaw in relying solely on population-level statistics for personal medical decision-making. Survival curves, medians, and CIs offer generalized probabilities, but the human brain - especially under stress - doesn’t process these figures as meaningfully predictive. They can feel abstract, cold, and ultimately unhelpful. A patient is not the average of 1408 pooled trial participants - they are a singular, complex biological system, with unique molecular, genetic, environmental, and psychological factors. In such cases, an N-of-1 narrative - where the focus is on this patient, this tumor, this biology - becomes not only more humane, but potentially more precise and strategic. Instead of guessing the odds from a collective outcome, we interrogate the individual’s tumor heterogeneity, MGMT methylation, mutation landscape, immune microenvironment, and therapeutic response using real-time, patient-derived data - such as single-cell transcriptomics or dynamic monitoring of circulating biomarkers.

This is not about disregarding clinical trials or statistics. It’s about translating them into actionable insight that centers the patient’s uniqueness, not their statistical resemblance to a broader cohort. So, when a patient asks, “Will this work for me?”, we owe them more than a percentage. We owe them a strategy - one that listens to their biology, adapts in real-time, and treats them not as a data point, but as the one case that matters most: Their own. But how can we work through standardized care frameworks and gravitate toward N-of-1? GBM remains one of the most treatment-resistant cancers, and response rates to current therapies are generally low[81]. Overall drug response rates in GBM remain modest, typically ranging from 10% to 40%, depending on the therapy[82] and patient-specific factors such as MGMT promoter methylation status. GBM’s inherent heterogeneity, invasive nature, and the protective blood-brain barrier contribute to its resistance to many treatments. Ongoing research is focused on identifying biomarkers to better predict treatment responses and developing novel therapies to improve outcomes for GBM patients. We have made significant contributions to GBM research, particularly through single-cell analyses. Our dedication to understanding GBM at the single-cell level aims to uncover mechanisms of tumor progression and resistance that could inform more effective therapeutic strategies.

Notable publications include: “Relapse pathway of GBM revealed by single-cell molecular analysis”[47]: This study employed scRNA-seq to examine primary and relapsed GBM tumors from a patient, identifying three mutated genes within individual cells involved in the RAS/GEF guanosine triphosphate-dependent signaling pathway - GBM’s lethal signaling pathway is revealed only through single-cell molecular analysis - not by bulk tumor profiling. The findings suggest that single-cell analysis can overcome the heterogeneity of bulk tumors, providing insights into subclonal evolution relevant to GBM relapse. GBM’s lethal behavior is fueled by its subclonal evolution and heterogeneity at the single-cell level. Our single-cell transcriptomic study uncovered relapse-associated mutations in the RAS/GEF guanosine triphosphate-dependent pathway, findings validated across three thousand patients. These results showed that single-cell molecular analysis can overcome the blind spots of bulk profiling, revealing relapse pathways critical for precision medicine. The hope scientists carry forward is clear: Single-cell cancer monitoring technologies hold the promise to identify relapse pathways in real time, personalize prognosis, and guide more effective therapies. “Cancer stem cells from a rare form of GBM involving the neurogenic ventricular wall”[83]: This research focused on isolating and characterizing cancer stem cells from a GBM tumor involving the lateral ventricles. The study demonstrated that CD133-positive cells from the tumor expressed various stem cell markers and exhibited tumor-forming capabilities both ex vivo and in vivo, suggesting their role in rapid tumor recurrence.

A scoping review and evidence map were conducted to explore optimal GBM treatment strategies - including combination regimens, dosing, and timing - with a focus on drug response rates to standard therapies in GBM patients, which show population-based patterns of responsiveness

TMZ^®: The standard chemotherapeutic agent for newly diagnosed GBM. TMZ exerts its cytotoxic effect by methylating guanine residues at the O6 position in DNA, which leads to mispairing during replication and triggers cell-cycle arrest and apoptosis[84]. However, the DNA repair protein MGMT counteracts this effect by directly reversing the methylation, thus conferring resistance to TMZ. MGMT expression is typically silenced by promoter methylation in about half of GBM tumors, but this correlation is not absolute - studies using patient-derived xenografts have shown that even some MGMT-unmethylated tumors may exhibit varying TMZ sensitivity depending on their capacity to dynamically regulate MGMT protein levels during treatment, which has implications for optimizing dosing regimens. Response rate: Approximately 30%-40% in newly diagnosed patients. MGMT promoter methylation: Patients with methylated MGMT promoters exhibit better responses to TMZ. This study indicated that MGMT promoter methylation was associated with significantly greater survival in GBM patients treated with TMZ and radiation; 2-year survival was 46% in patients with tumor MGMT hypermethylation vs 14% in patients with MGMT hypomethylation.

Bevacizumab (Avastin): An anti-vascular endothelial growth factor antibody used primarily in recurrent GBM. Objective response rate: A meta-analysis reported a 6% complete response and a 49% partial response in recurrent GBM patients treated with bevacizumab (BEV)[85]. OS: While BEV improves PFS, studies have not demonstrated a significant benefit in OS for recurrent GBM (rGBM) patients[86]. BEV therapy in rGBM unravels: (1) PFS, not OS: BEV improves PFS and offers palliative and cognitive benefits, but does not demonstrate a consistent OS advantage in recurrent GBM based on current evidence; (2) Combination therapy outperforms monotherapy: BEV shows greater efficacy when combined with agents like lomustine or radiotherapy, rather than used alone, suggesting that combination regimens may be more effective in prolonging survival; (3) Patient selection matters: Clinical and molecular markers - including IDH mutation status, large tumor burden, double-positive imaging signs, and low ADC values - may help predict which rGBM patients will respond best to BEV; (4) Lower doses may suffice: Low-dose BEV appears to be as effective as standard dosing, which could help reduce treatment-related side effects and improve patient tolerability; and (5) Timing and tailoring still needed: The optimal timing (window of opportunity) for initiating BEV therapy remains unclear, and high-quality studies are needed to refine BEV-based treatment strategies and better define responsive subpopulations.

Tumor treating fields: A non-invasive therapy that uses low-intensity, intermediate-frequency alternating electric fields to disrupt cancer cell division, particularly targeting mitosis in GBM cells[87]. Efficacy: Tumor treating fields (TTF), in combination with TMZ, have been shown to improve PFS and OS in newly diagnosed GBM patients[88]. However, specific response rates vary, and further studies are ongoing to better define its effectiveness. The therapy is delivered through a wearable, portable device (Optune^®) that generates electric fields via skin-applied transducer arrays, and it has received the United States Food and Drug Administration approval for both newly diagnosed and recurrent GBM[89]. TTF is generally well tolerated, with skin irritation being the most commonly reported side effect, and it selectively targets tumor cells while sparing normal tissue[90]. Ongoing clinical trials in exploring the combination of TTF with radiotherapy, chemotherapy, and immunotherapy across a broader range of cancers, with active research into its mechanisms and expanded indications, offer promising potential for managing GBM, especially given the tumor’s resistance due to its complex microenvironment. It emphasizes the growing trend of “cocktail therapy”, where TTFs are integrated with chemotherapy, immunotherapy, and emerging technologies to enhance therapeutic outcomes[91]. For instance, a phase 3, randomized clinical trial (NCT04471844) in Poland plans to enroll 950 newly diagnosed GBM patients to assess whether initiating TTF during concurrent chemoradiation improves survival outcomes[92]. Additionally, the TIGER trial (NCT03258021) in Germany aims to include 710 patients to evaluate TTF as part of routine clinical practice [on the planned subgroup analysis (MGMT and age)][93]. These studies reflect a broader effort to investigate TTFields alongside chemotherapy, radiation, and immunotherapy.

ICI (e.g., nivolumab): These therapies aim to enhance the body’s immune response against tumor cells. Response rate for nivolumab and for BEV was unraveled[94]. In clinical scenarios, the main numerical findings from the CheckMate 143 phase 3 trial comparing nivolumab (n = 184) to BEV (n = 185) in recurrent GBM patients (n = 369): (1) Median OS was comparable between the two groups: 9.8 months for nivolumab vs 10.0 months for BEV (hazard ratio = 1.04; P = 0.76), indicating no survival benefit with PD-1 blockade; (2) The 12-month OS rate was 42% in both arms, reinforcing the lack of significant difference in long-term survival outcomes; (3) Objective response rate was substantially higher with BEV (23.1%) compared to nivolumab (7.8%), showing BEV’s capacity for measurable tumor shrinkage; (4) MGMT promoter methylation was reported in about 23% of patients in both groups, with similar distributions of unmethylated and unknown cases across the arms; and (5) Grade 3/4 treatment-related adverse events occurred at similar rates: 18.1% nivolumab and 15.2% for BEV, with no unexpected neurological toxicities or treatment-related deaths observed in either group.

Antibody block mutations: In the course of antibody affinity maturation, germinal centre (GC) B cells mutate their immunoglobulin heavy- and light-chain genes in a process known as somatic hypermutation (SHM). The study “Transient silencing of hypermutation preserves B cell affinity during clonal bursting” uncovers a safeguard in GC biology: Strategic suppression of SHM during intense clonal expansion[95]. By combining intravital imaging, image-based sorting, and mathematical modeling, the authors reveal that B cells in rapid proliferative bursts bypass the G0-like phase - where SHM typically occurs - thereby protecting antigen-binding affinity from deleterious mutations. This work refines the understanding of GC dynamics, illustrating how the immune system balances diversification with the preservation of high-affinity clones during selection gaps.

FOSTERING DIALOGUE FOR FUTURE N-OF-1 TREATMENT

Current precision medicine in cancer is incomplete and ineffective for many patient subgroups due to its narrow genetic focus and exclusion of critical physiological variables. A truly personalized precision medicine must move beyond oncogenic drivers to encompass tumor-immune interactions, pharmacogenomic variations, and preventative strategies, ensuring that treatment is not only molecularly precise but holistically tailored to each patient’s unique disease trajectory. Nam et al’s work[1] extends beyond its dataset, catalyzing dialogue, hypothesis generation, and collaborative research. To unravel the complexities of SCC prognosis and treatment, researchers must adopt a multidisciplinary perspective, such as spatial and temporal switch between dominating and dormant subclones[73], tumor microenvironment[96], tissue elasticity[97] and the single-cell molecular profiling with an AI perspective[16], to gravitate toward prevention. This approach integrates molecular discoveries with clinical insights, paving the way for innovative strategies in subclonal evolution of cancer progression and post-treatment recurrence through subclonal switchboard signaling, with important implications for personalized therapy development, i.e., aimed at the innovative framework for N-of-1 trials of individualized gene-targeted therapies of genetic mutations[98].

CONCLUSION

In conclusion, that study identifies an unexpected association between p16 positivity and poor prognosis in HNCSCC, underscores the dual and mutation-dependent roles of p53, and highlights the limitations of current HPV detection strategies. Addressing these complexities requires not only methodological refinement but also the adoption of multi-omics and spatial single-cell approaches to unravel the interactions within the microenvironment. As these technologies mature, they promise to transform biomarker discovery from static indicators into dynamic, patient-tailored predictors of recurrence, metastasis, and treatment response. We envision a future where such integrative frameworks enable clinicians to stratify risk and design therapies with precision - shifting from one-size-fits-all protocols to truly individualized care - including: (1) Unexpected poor prognosis link of p16 in HNCSCC; (2) The duality of p53 mutations; (3) Methodological refinements for HPV detection; (4) How multi-omics, spatial omics, and N-of-1 approaches will enable precision risk stratification and tailored therapy; and (5) We suggest kinetic and microenvironmental models - such as “load-and-lock” barriers and stabilization of dormancy or immune synapses - as hypothesis-generating strategies to interrupt clonal transitions and refine outcome prediction. Collectively, these perspectives call for redefined biomarker frameworks in HNCSCC and advocate ethically inclusive, mechanism-driven studies that link molecular discovery with truly individualized patient care. To truly remedy the cancer problem, we must confront its root cause rather than merely treating the visible symptom - a growing tumor - with surgery, chemotherapy, or immunotherapy. The relentless expansion of “populist” cells is only a manifestation. The deeper question remains: What is the root cause of the anxiety that pervades cancer management? When the immune system is broken - driven by uncontrolled subclonal evolution of malignant cells - cancer becomes a wound that never heals, a failure of the body’s intrinsic self-healing and self-salvage system. In such a state, cancer cell “populists” inevitably pursue their advantage, sooner or later. Halting cancer at the level of a single cell is not only possible but essential for true prevention.