Comprehensive geriatric assessment guiding treatment and survival in elderly digestive tumor patients: A retrospective study

Abstract

BACKGROUND

The intensification of global aging has led to an increasing burden of digestive system tumors in the elderly. The comprehensive geriatric assessment (CGA) provides important basis for developing individualized treatment plans by systematically assessing factors such as patient function, nutrition, and comorbidities.

AIM

To examine the value of the CGA in treatment decision-making and survival prognosis among elderly patients with digestive system tumors.

METHODS

This retrospective study analyzed the clinical data of 210 elderly patients with digestive system tumors admitted between April 2022 and July 2025. Of these, 108 patients who received traditional diagnostic and therapeutic decisions (April 2022 to December 2023) constituted the control group, whereas 102 patients who received CGA-guided diagnostic and therapeutic decisions (January 2024 to July 2025) comprised the observation group. The initial treatment strategy, intensity, safety, and tolerance, as well as progression-free survival (PFS), were compared between the two groups.

RESULTS

Fewer patients in the observation group received standard treatment and standard-intensity chemotherapy compared with the control group, whereas more received reduced-intensity chemotherapy. The overall incidence of adverse reactions (≥ grade 3) was 28.4% in the observation group, compared with 45.4% in the control group. Treatment delay and discontinuation rates were lower in the observation group (20.6% and 11.8%, respectively) than in the control group (32.4% and 24.1%, respectively). The treatment completion rate was higher in the observation group (88.2%) than in the control group (75.9%). The median PFS was longer in the observation group (11.5 months, 95% confidence interval: 10.2-12.8) than in the control group (9.1 months, 95% confidence interval: 8.0-10.2). All differences were statistically significant (P < 0.05). Multivariate Cox regression analysis showed that a CGA-assessed “frail” status was an independent risk factor affecting patient PFS (P < 0.05).

CONCLUSION

CGA-guided diagnostic and therapeutic strategies facilitate precise, individualized decision-making for elderly patients with digestive system tumors. By tailoring treatment selection and intensity to overall health status, this approach optimizes the balance between efficacy and safety and is associated with improved PFS, supporting its clinical utility.

Key Words: Comprehensive geriatric assessment; Digestive system neoplasms; Treatment decision-making; Survival prognosis; Retrospective studies

Core Tip: Comprehensive geriatric assessment plays a pivotal role in the clinical management of elderly patients with digestive system tumors. By systematically evaluating their physical function, cognitive status, nutritional condition and social support, comprehensive geriatric assessment precisely guides the formulation of personalized treatment regimens. Notably, this evidence-based approach effectively mitigates treatment-related toxicities, enhances patients’ treatment tolerance, and extends their progression-free survival, thus exerting remarkable clinical benefits and practical value in geriatric oncology practice.

INTRODUCTION

With the accelerating trend of global population aging, the incidence of malignant tumors shows a significant positive correlation with age, making advanced age the strongest risk factor[1,2]. Digestive system tumors, such as esophageal, gastric, colorectal, liver, and pancreatic cancers, are among the most common malignancies in China, ranking high in incidence and mortality rates, particularly burdening the elderly population[3,4]. Traditional oncology decision-making primarily relies on factors such as tumor-node-metastasis (TNM) stage, pathological type, and molecular markers of the tumor. However, it often overlooks the impact of geriatric syndromes, such as the patient's overall functional status, comorbidities, cognition, psychological state, and social support. This oversight can lead to overly conservative approaches - missing opportunities for active treatment and prolonged survival - or to excessive treatment that causes unnecessary toxic side effects, impairs quality of life, or accelerates death[5,6]. Therefore, accurate assessment of the overall health status of elderly patients with cancer, along with the development of individualized, optimized treatment strategies, has become a core challenge and a cutting-edge topic in geriatric oncology.

Recently, comprehensive geriatric assessment (CGA), as a multidimensional tool, has gradually shown unique value in diagnosing and treating elderly patients with cancer[7,8]. By systematically analyzing information regarding the patient’s physical function, mental health, social support, and comorbidities, CGA can comprehensively reflect the overall health status of elderly patients. Thus, in this retrospective study we analyzed the clinical data of 210 elderly patients with digestive system tumors admitted to our hospital between April 2022 and July 2025. The study aimed to examine the role of CGA in treatment decision-making and survival prognosis, thereby providing a scientific basis for optimizing individualized treatment.

MATERIALS AND METHODS

Participants

This study retrospectively analyzed clinical data of 210 elderly patients with digestive system tumors admitted to our hospital from April 2022 to July 2025. Of these patients, 108 admitted from April 2022 to December 2023 who received traditional oncology decision-making comprised the control group, whereas 102 patients admitted from January 2024 to July 2025 who received CGA-guided oncology decision-making constituted the observation group. The study protocol was approved by the hospital ethics committee.

Inclusion criteria were as follows: (1) Age ≥ 65 years; (2) Diagnosis of a digestive system malignant tumor by histopathology, including gastric, colorectal, esophageal, liver, pancreatic cancers, etc.; (3) Complete clinical, treatment, and follow-up records; (4) Eastern Cooperative Oncology Group (ECOG) performance status score ≤ 2 at initial diagnosis; and (5) Informed consent obtained from patients and their families for the diagnostic and therapeutic procedures.

Exclusion criteria were as follows: (1) Concurrent primary malignant tumors in other organ systems; (2) Substantially incomplete clinical data; (3) Inability to cooperate with relevant assessments due to severe mental illness or cognitive impairment; (4) Expected survival of less than 3 months (terminal stage); and (5) Loss to follow-up.

Baseline demographic and clinical characteristics were balanced between the groups. No statistically significant differences were observed in age, sex distribution, tumor type, or ECOG performance status (all P > 0.05), indicating good comparability between the cohorts (Table 1).

Table 1 Comparison of baseline characteristics between the two groups, n (%).

Group	n	Age (year), mean ± SD	Gender		Tumor type					ECOG grade (point)
			Male	Female	Colorectal cancer	Gastric cancer	Esophageal cancer	Liver cancer	Pancreatic cancer
Observation group	102	76.5 ± 4.8	59 (57.8)	43 (42.2)	46 (45.1)	28 (27.5)	10 (9.8)	12 (11.1)	6 (5.9)	1.2 ± 0.5
Control group	108	77.1 ± 5.2	65 (60.2)	43 (39.8)	48 (44.4)	32 (29.6)	12 (11.1)	10 (9.8)	6 (5.6)	1.1 ± 0.4
χ2/t value	-	0.851	0.102		1.234					1.666
P value	-	0.396	0.749		0.872					0.107

ECOG: Eastern Cooperative Oncology Group.

Methods

Control group: Patients in the control group were treated based on traditional oncology decision-making: (1) After admission, the attending physician collected medical history, performed physical examination, and completed imaging studies (computed tomography, magnetic resonance imaging, positron emission tomography-computed tomography) and laboratory tests to diagnose the tumor and determine its stage; (2) A multidisciplinary team, comprising medical oncologists, surgeons, and radiation oncologists, discussed the case. Based on comprehensive consideration of tumor factors (TNM stage, pathological type, differentiation degree, molecular markers like Kirsten rat sarcoma viral oncogene homolog/neuroblastoma rat sarcoma viral oncogene homolog/B-Raf proto-oncogene, serine/threonine kinase, human epidermal growth factor receptor 2, microsatellite instability) and general patient factors (age, ECOG score, routine blood test results, liver and kidney function, etc.), a treatment plan - such as surgery, chemotherapy, radiotherapy, targeted therapy, or immunotherapy - was formulated; and (3) During treatment, the plan was adjusted promptly based on the patient’s tolerance and short-term efficacy. In case of severe adverse reactions or disease progression, the treatment strategy was re-evaluated and adjusted.

Observation group: Patients in the observation group underwent treatment guided by CGA-based oncology decision-making: (1) After admission, in addition to routine examinations, a unified CGA was performed by trained geriatricians and oncology nurses, covering eight domains: Comorbidities, functional status, nutritional status, cognitive function, psychological state, frailty, polypharmacy, and social support; (2) According to the CGA results, patients were categorized into three groups: “fit” (no significant or only minor, manageable issues across CGA domains; expected to tolerate standard treatment), “vulnerable” (significant impairment in one or more domains, with some issues that can be improved through intervention; high risk for complications with standard treatment, necessitating adjustment), and “frail” (severe dependency, severe frailty, multiple irreversible geriatric syndromes, or inability to tolerate any form of active antitumor therapy); (3) After integrating oncological factors, general patient factors, and CGA results, multidisciplinary team, comprising medical oncologists, surgeons, radiation oncologists, geriatricians, and oncology nurses, held a discussion formulated an individualized treatment plan according to the “Treatment Decision Support Framework Guided by CGA Classification”; and (4) Patient indicators were dynamically monitored during treatment, and adjustments were made in real time based on CGA assessment results. CGA was re-performed, and the treatment plan was adjusted accordingly if significant functional decline or severe adverse reactions occurred. Details of the CGA domain criteria and the treatment decision support framework (examples) are shown in Tables 2 and 3.

Table 2 Comprehensive geriatric assessment dimension assessment tools, abnormality criteria, and general intervention measures.

CGA assessment dimensions	Assessment tool/method	Abnormality threshold	General interventions
Comorbidities	CCI	CCI ≥ 2	Refer to geriatrics or relevant specialty; optimize comorbidity management
Functional status	Activities of Daily Living scale	Total score < 100	Refer to rehabilitation medicine; devise a rehab plan; arrange home-care services
Nutritional status	MNA-SF	≤ 11 points	Refer to clinical nutrition; provide dietary counseling, oral nutritional supplements or mandatory nutritional support
Cognitive function	MMSE	Illiterate ≤ 17; primary-school ≤ 20; middle-school ≤ 24	Enhance communication with primary caregiver; remind and supervise medication; schedule cognitive training
Psychological status	GDS-15	≥ 5 points	Refer to psychiatry/psychology; provide psychological counseling; strengthen family emotional support
Frailty status	CFS	Grade ≥ 4	Use as core input for MDT decisions; initiate comprehensive nutrition, rehabilitation and supportive care
Polypharmacy	Medication list review	Concurrent use of ≥ 5 drugs	Clinical pharmacist or geriatrician conducts medication reconciliation; discontinue non-essential or inappropriate drugs
Social support	Structured interview	Living alone/no stable caregiver/financial hardship	Social-worker intervention; assist with subsidy applications; link to community support services

CGA: Comprehensive geriatric assessment; CCI: Charlson comorbidity index; MNA-SF: Mini Nutritional Assessment-Short Form; MMSE: Mini-Mental State Examination; GDS-15: Geriatric Depression Scale-15; CFS: Clinical Frailty Scale; MDT: Multidisciplinary team.

Table 3 Treatment decision support framework under comprehensive geriatric assessment classification guidance (example in locally advanced stage III gastric cancer).

CGA grade	Target	Core treatment strategy	Specific example
Health	Cure	Standard neoadjuvant/adjuvant chemotherapy + radical surgery	Perioperative chemotherapy: Standard FLOT regimen (doxorubicin + oxaliplatin + fluorouracil/Leucovorin) is administered.rmed
			Surgery: Radical gastrectomy (D2 standard) is perfo
			Postoperative care: Complete all planned adjuvant chemotherapy cycles
Fragile	Control and function equally important	Optimize the standard protocol to reduce treatment-related risks and ensure successful completion of treatment	Pretreatment support: Initiate intensive nutritional support prior to chemotherapy and administer G-CSF prophylactically
			Perioperative chemotherapy: Employ the less toxic SOX (tegafur + oxaliplatin) or XELOX (capecitabine + oxaliplatin) combination regimens, with initial doses reduced to 80% of standard levels
			Surgery: Perform radical resection, followed by adjustment of adjuvant chemotherapy based on postoperative recovery
Caducity	Quality of life	Avoid high-intensity radical treatment and focus on controlling tumor-related symptoms, maintaining the feeding channel and quality of life	Nonsurgical local treatment: Palliative radiotherapy to control bleeding or pain
			Systemic treatment: Monotherapy (e.g., tegafur) or targeted therapy (e.g., trastuzumab for HER2-positive patients)
			Core measures: Enhanced nutritional support; professional pain and symptom management; if obstruction exists, perform gastrojejunostomy or place intestinal stent

CGA: Comprehensive geriatric assessment; FLOT: Folinic acid + leucovorin + oxaliplatin + docetaxel; G-CSF: Granulocyte colony-stimulating factor; SOX: S-1 + oxaliplatin; XELOX: Capecitabine + oxaliplatin.

Follow-up: All patients were followed up regularly until the cutoff date of October 15, 2025.

Observation indicators

Differences in initial treatment strategy, treatment intensity, safety, and tolerability, as well as survival prognosis indicators, were compared between the two groups.

Initial treatment strategy: The initial treatment modalities received by patients in both groups were recorded and categorized primarily into standard therapy (curative surgery + adjuvant chemotherapy), palliative therapy (palliative chemotherapy/targeted therapy/immunotherapy), or supportive care only.

Treatment intensity: Treatment intensity was evaluated using the relative dose intensity (RDI) of chemotherapeutic drugs, calculated as RDI = (actual dose/planned dose) × 100%. RDI ≥ 85% was defined as standard intensity, 60%-85% as reduced intensity, and < 60% as low intensity.

Treatment safety: According to the Common Terminology Criteria for Adverse Events version 5.0, the occurrence of all chemotherapy-related adverse reactions of grade ≥ 3 (including nausea or vomiting, diarrhea, myelosuppression, liver or kidney dysfunction, neurotoxicity, etc.) during treatment was recorded for groups.

Treatment tolerability: The treatment delay (chemotherapy delayed > 7 days due to toxicity), discontinuation (permanent discontinuation of chemotherapy because of toxicity), and completion (completion of planned chemotherapy cycles) rates were recorded for both groups.

Survival prognosis indicator: Progression-free survival (PFS) was used for evaluation, and the impact of CGA classification on PFS was analyzed.

Statistical analysis

Data were analyzed using SPSS software (version 26.0). Data conforming to a normal distribution according to the Shapiro-Wilk test were expressed as mean ± SD and compared between groups using the t-test. Count data were expressed as n (%) and compared between groups using the χ² or Fisher’s exact test. Survival curves were plotted using the Kaplan-Meier method and compared using the Log-rank test. Multivariate survival analysis was performed using the Cox proportional hazards regression model. A P value < 0.05 was considered statistically significant.

RESULTS

Comparison of treatment decisions and intensity between the two groups

Significant differences were observed between the two groups regarding initial treatment strategies and chemotherapy intensity (Table 4). The proportion of patients receiving standard treatment was significantly lower in the observation group compared with the control group (35.3% vs 48.1%, P = 0.035), whereas a higher proportion of patients in the observation group received palliative therapy (46.1% vs 36.1%). Regarding treatment intensity, fewer patients in the observation group received standard-intensity chemotherapy compared with the control group (41.2% vs 60.2%, P = 0.006), whereas the proportion of patients receiving reduced-intensity chemotherapy was significantly higher in the observation group (47.1% vs 29.6%, P = 0.008) compared with the control group. No significant difference was found in the proportion of patients receiving low-intensity chemotherapy between the two groups (P = 0.709). These findings indicate that CGA-guided decision-making was associated with more frequent adjustment of treatment intensity while maintaining individualized treatment strategies (Table 4).

Table 4 Comparison of treatment decisions and intensity between the two groups, n (%).

Group	n	Initial treatment strategy			Treatment intensity
		Standard treatment	Palliative therapy	Supportive care only	Normal intensity	Reduced intensity	Low intensity
Observation group	102	35 (35.3)	48 (46.1)	19 (18.6)	42 (41.2)	48 (47.1)	12 (11.8)
Control group	108	53 (48.1)	38 (36.1)	17 (15.7)	65 (60.2)	32 (29.6)	11 (10.2)
χ2/t value	-	4.427	2.686	0.332	7.699	6.941	0.139
P value	-	0.035	0.101	0.564	0.006	0.008	0.709

Comparison of treatment safety between the two groups

The overall incidence of grade ≥ 3 treatment-related adverse events was significantly lower in the observation group (28.4%) than in the control group (45.4%; P = 0.009; Table 5). Specifically, the observation group demonstrated reduced rates of severe neutropenia (11.8% vs 23.1%), thrombocytopenia (7.8% vs 13.9%), diarrhea (4.9% vs 10.2%), and neurotoxicity (3.9% vs 7.4%) compared with the control group. Although individual adverse events varied in frequency, the overall reduction in severe toxicity was statistically significant, indicating improvement in safety profiles under CGA-guided management.

Table 5 Comparison of treatment safety between the two groups, n (%).

Group	n	Neutropenia	Thrombocytopenia	Diarrhea	Neurotoxicity	Overall grade ≥ 3 adverse events
Observation group	102	12 (11.8)	8 (7.8)	5 (4.9)	4 (3.9)	29 (28.4)
Control group	108	25 (23.1)	15 (13.9)	11 (10.2)	8 (7.4)	49 (45.4)
χ2/t value	-	-	-	-	-	6.834
P value	-	-	-	-	-	0.009

Comparison of treatment tolerability between the two groups

Rates of treatment delay (20.6%) and discontinuation (11.8%) were lower in the observation group than in the control group (32.4% and 24.1%, respectively), whereas the completion rate was higher (88.2% vs 75.9%). The differences were statistically significant (P < 0.05; Table 6).

Table 6 Comparison of treatment tolerability between the two groups, n (%).

Group	n	Treatment delay	Treatment discontinuation	Treatment completion
Observation group	102	21 (20.6)	12 (11.8)	90 (88.2)
Control group	108	35 (32.4)	26 (24.1)	82 (75.9)
χ2/t value	-	3.945	5.678	5.512
P value	-	0.047	0.017	0.019

Kaplan-Meier curve analysis of PFS in the two groups

The median PFS was significantly longer in the observation group (11.5 months, 95% confidence interval: 10.2-12.8) than in the control group (9.1 months, 95% confidence interval: 8.0-10.2). The difference was statistically significant (Log-rank χ² = 4.12, P = 0.042; Figure 1).

Figure 1 Kaplan-Meier curve analysis of progression-free survival in both groups. CGA: Comprehensive geriatric assessment; PFS: Progression-free survival; CI: Confidence interval.

Multivariate Cox regression analysis of PFS in the two groups

Multivariate Cox regression, adjusted for age, ECOG score, and TNM stage, showed that a CGA-assessed “vulnerable” status was an independent risk factor affecting patient PFS (P < 0.05; Table 7).

Table 7 Multivariate Cox regression analysis of progression-free survival in both groups.

Variable	HR	95%CI	P value
Age (per five-year increase)	1.32	1.05-1.66	0.018
ECOG score (≥ 2 vs < 2)	1.87	1.24-2.82	0.003
TNM stage (stage IV vs I-III)	2.45	1.65-3.64	< 0.001
Comprehensive geriatric assessment category (ref = control group)	-	-	0.028
Fit	0.65	0.42-1.01	0.055
Vulnerable	0.71	0.50-0.99	0.046
Frail	1.05	0.62-1.78	0.856

CI: Confidence interval; HR: Hazard ratio; ECOG: Eastern Cooperative Oncology Group; TNM: Tumor-node-metastasis.

DISCUSSION

Elderly patients with cancer represent a highly heterogeneous population in whom physiological age often diverges from chronological age. The prevalence of geriatric syndromes such as comorbidities, cognitive impairment, functional decline, malnutrition, polypharmacy, and psychosocial and social support issues make diagnostic and therapeutic decision-making for this group highly challenging[9,10]. CGA is a multidimensional, interdisciplinary tool that provides clinicians with precise “profiles” of elderly patients with cancer by evaluating multiple domains, including physical function, nutritional status, cognitive and psychological function, and social support[11,12]. Studies have shown that the application of CGA in breast, lung, urological cancers, as well as other cancers, can predict treatment complications, toxicities, and survival outcomes, and guide treatment adjustments[13,14]. Recent updates from the International Society of Geriatric Oncology strongly recommend CGA as a standard component of oncologic assessment for elderly people[15,16]. Accordingly, our institution began implementing CGA-guided oncology decision-making for elderly patients with digestive system tumors in January 2024, achieving relatively ideal results.

Our findings showed that, compared with the control group, the observation group had a significantly lower proportion of patients receiving standard therapy and standard-intensity chemotherapy (P < 0.05), and a higher proportion receiving reduced-intensity chemotherapy (P < 0.05). Traditional oncology decision-making - based primarily on biological indicators such as TNM stage and pathological type - has inherent limitations in elderly patients whose physiological reserve is diminished and comorbidities are complex. Such approaches may overestimate tolerance in “seemingly healthy” elderly patients, leading to overtreatment and severe toxicity[17,18], while underestimating the potential for active treatment in “vulnerable” patients whose condition could improve with appropriate intervention, resulting in undertreatment[19,20].

CGA systematically evaluates multiple domains - comorbidities, function, nutrition, cognition, psychology, frailty, medication use, and social support - to refine the distinction between physiological age and health status, categorizing patients into “fit”, “vulnerable”, and “frail”[21,22]. Standard or near-standard regimens can still be considered in patients identified as “fit” despite advanced age, enabling curative or optimal disease control while avoiding unnecessary undertreatment based on age alone[23,24]. Conversely, “vulnerable” patients possess limited physiological reserve and face high risks from unmodified standard therapy. For these individuals, the CGA-guided strategy in this study employed reduced-intensity chemotherapy and enhanced supportive care, successfully balancing tumor control with maintenance of function and quality of life. This aligns with the findings of Bonneau et al[25], who reported that CGA altered nearly one-third of treatment decisions in elderly prostate cancer patients, most often by reducing treatment intensity or increasing supportive care. For frail patients, our decision framework recommended avoiding high-intensity curative treatments and instead emphasizing comfort and quality of life - reflecting the geriatric oncology principle that “sometimes, no treatment is the best treatment”. For those with very limited life expectancy, transitioning to a palliative care model focused on symptom relief and psychosocial support represents a humane and patient-centered approach.

In addition, the incidence of grade ≥ 3 adverse events, treatment delays, and treatment discontinuations was lower in the observation group than in the control group (P < 0.05), while the treatment completion rate was higher (P < 0.05). Elderly patients with digestive system tumors frequently experience polypharmacy, malnutrition, and mood disorders, all of which impair tolerance to systemic therapy. The CGA-guided model enables early identification of risk factors such as frailty, nutritional deficiency, and functional dependence, allowing clinicians to anticipate vulnerability and implement targeted interventions before initiating treatment. By integrating anti-cancer therapy with concurrent geriatric interventions, this model improves patients’ overall physiological and psychological resilience, reduces complications, and facilitates completion of planned therapy[26]. A systematic review by Disalvo et al[27] similarly concluded that CGA enhances treatment completion rates and reduces toxicity in elderly patients receiving systemic anti-cancer therapy, consistent with our findings.

The median PFS was also significantly longer in the observation group than in the control group (P < 0.05), and a CGA-assessed vulnerable status emerged as an independent risk factor for PFS (P < 0.05). Although treatment intensity appeared “milder” in the observation group, their median PFS was longer. This seemingly paradoxical result demonstrates that CGA-guided decision-making helps clinicians allocate effective treatment precisely to patients who can benefit from it and tolerate it, profoundly revealing the role of CGA in optimizing resource allocation and maximizing treatment benefit. The “vulnerable” status identified by CGA reflects the depletion of physiological reserve and diminished ability to cope with stress, which affects treatment tolerance and may be intrinsically linked to tumor biology and the antitumor immunity of the body. Therefore, implementing active interventions such as physical exercise, nutritional support, and comorbidity management targeted at the “vulnerable” status is beneficial for quality of life and has potential antitumor and prognostic improvement significance. This echoes the conclusions of studies by Yamashita et al[28] in esophageal cancer and Itami et al[29] in gastric cancer.

This study has limitations. The follow-up duration was relatively short, and the sample size was small, which may introduce selection and information bias. Future large-scale, multicenter, prospective randomized controlled trials are warranted to comprehensively validate the long-term efficacy and safety of CGA-guided oncology decision-making for elderly patients with digestive system tumors.

CONCLUSION

The application of CGA-based diagnosis and treatment decisions in elderly patients with digestive system tumors helps guide the formulation of more individualized treatment plans. This approach effectively reduces treatment toxicity, improves treatment tolerance, and prolongs PFS, while ensuring efficacy; hence, is worthy of application.