INTRODUCTION
Biliary tract cancer (BTC) is a highly heterogeneous disease, including cholangiocarcinoma (CCA) and gallbladder carcinoma (GBC)[1]. BTC accounts for about 3% of all gastrointestinal tumors, among which GBC is the most common, accounting for 80%–95% of BTC, and ranks the sixth in the global incidence of gastrointestinal tumors[2]. At present, the global incidence of BTC is on the rise, and is highest in Asian countries[3]. Most BTC is adenocarcinoma (approximately 95%), highly aggressive, and found at advanced stages. The prognosis is extremely poor, and the 5-year survival rate is < 5%[4]. So far, radical treatment of BTC relies on complete tumor resection[5]. Therefore, early diagnosis and accurate evaluation are important for treatment decisions and prognosis of BTC.
At present, histopathological and cytological examinations are the gold standard for diagnosis of BTC, while imaging examinations are helpful for comprehensive evaluation of patients[6]. Traditional imaging techniques, which include computed tomography (CT) and magnetic resonance imaging (MRI), are capable of disclosing information about lesion characteristics that are relevant to disease staging, treatment planning, and prognostic evaluation. This information encompasses tumor attributes, vascular infiltration, lymph node (LN) affection, and distant spread. During the initial diagnostic phase, ultrasound is primarily utilized as a screening instrument. Conversely, CT and MRI scans are crucial for evaluating tumor features and their anatomical relationship with adjacent structures, as well as for the detection of distant metastases[2]. Nevertheless, conventional imaging modalities such as these have limitations; they cannot depict lesion activity or the overall systemic status, and they face certain constraints in diagnosing non-enlarged metastatic LNs and tumor recurrence. The 18F-fludeoxyglucose (FDG) positron emission tomography (PET) CT (18F-FDG PET/CT) combines anatomical and functional imaging and is widely used for preoperative and postoperative evaluation of cancer[7].
This review aims to summarize the research progress of 18F-FDG PET/CT in BTC over the past decade (especially in the past 3 years), focusing on its unique value in clinical staging, response evaluation, prognosis prediction, and pathological information analysis, while comparing its advantages and limitations with traditional imaging techniques. Given the distinct epidemiological, etiological, biological features, and imaging appearances of biliary tract tumors based on their location and type, this paper undertakes a separate review of each.
18F-FDG PET/CT MANIFESTATIONS AND DIAGNOSTIC DEFECTS OF BTC
BTC
BTC is an aggressive tumor originating from the gallbladder or bile duct, which is characterized by early lymphatic and hematogenous metastasis[8]. The 18F-FDG PET/CT has certain diagnostic value for primary BTC. The sensitivity of 18F-FDG PET/CT in diagnosing BTC may be related to the expression of glucose transporter (GLUT) proteins in biliary tract epithelial cells. Specifically, GLUT-1 and GLUT-3 act as the main facilitators of cellular glucose uptake, demonstrating a high affinity for glucose and playing an essential role in metabolic pathways. The meta-analysis of Lamarca et al[7] included a total of 30 studies, which collectively reported systematic review data on the assessment of primary tumors in BTC patients using 18F-FDG PET/CT. The pooled results indicated a sensitivity of 91.7% and a specificity of 51.3%. At the same time, the study performed a comparative analysis of the maximum standardized uptake value (SUVmax) data from cancer patients versus those without cancer. The cancer patients exhibited a significantly higher SUVmax compared to noncancer patients (6.2 vs 2.8). A retrospective study was conducted on 99 patients with suspected BTC to compare the assessment of primary lesions using 18F-FDG PET/CT and multidetector CT (MDCT), against the final pathological or clinical diagnosis. The sensitivity, specificity, positive predictive value, negative predictive value, and diagnostic accuracy of 18F-FDG PET/CT and MDCT were 90.2%, 70.6%, 93.7%, 60.0%, 86.9%, 84.2%, 70.6%, 93.2%, 48.0%, and 81.8%, respectively. The 18F-FDG PET/CT had high diagnostic value for the primary tumor of BTC, but compared with CT, 18F-FDG PET/CT had no obvious advantage[9]. The study also found that among different BTC subtypes, the highest median SUVmax was found for intrahepatic cholangiocarcinoma (ICC) (9.4%), followed by GBC (8.4%), and extrahepatic cholangiocarcinoma (ECC) (4.9%).
For LN metastasis of BTC, CT and MRI are mainly based on the size of LNs, which is prone to produce false-positive or false-negative results. The 18F-FDG PET/CT improves the detection rate of LN metastases by integrating both anatomical and metabolic information. Lee et al[9] conducted a comparison between 18F-FDG PET/CT and MDCT for assessing regional LNs, revealing that 18F-FDG PET/CT had a higher positive predictive value of 94.1% for diagnosing regional LN metastasis compared to 77.5% for MDCT. A meta-analysis in 2019 also showed that, in terms of LN metastasis, the combined sensitivity and specificity of 18F-FDG PET/CT were 88.4% and 69.1%[7], which had better diagnostic advantages compared with MDCT.
The 18F-FDG PET/CT exhibits considerable value in identifying unexpected distant metastases that are not detected by traditional imaging techniques. Petrowsky et al[10] included 61 patients with BTC. All distant metastases were detected by 18F-FDG PET/CT (12/12), whereas only three were identified by contrast-enhanced CT. In a retrospective analysis of patients with BTC who underwent both 18F-FDG PET/CT and MDCT, distant metastases were observed in 19 of 82 patients (23.2%). The sensitivity of 18F-FDG PET/CT (94.7%) was significantly higher than that of MDCT (63.2%). This study showed that compared with MDCT, 18F-FDG PET/CT has significant advantages in detecting distant metastasis of biliary tract malignant tumors[9].
Common reasons that interfere with the diagnosis of 18F-FDG PET/CT for BTC include false-positive results related to bile duct infection, coexistence of biliary stents, and various benign diseases[11,12]. The bile ducts of normal healthy individuals do not show FDG uptake; however, causes of bile duct abscess and inflammation such as sclerosing cholangitis, cholangitis, and cholecystitis may lead to false-positive 18F-FDG PET/CT interpretation; however, they may also mask underlying malignant lesions. Benign diseases such as tuberculosis, Langerhans cell histiocytosis, and adenomyomatosis also increase the challenge of 18F-FDG PET/CT interpretation[13]. The detection of small lesions by 18F-FDG PET/CT may be hindered by partial volume effects, artifacts related to respiratory motion, and elevated activity in the adjacent liver and intestine[14].
CCA
CCA is a malignant tumor originating from the bile duct epithelium and its surrounding glands. According to the anatomical location, CCA can be divided into ICC and ECC. ECC is divided into perihilar cholangiocarcinoma (PCC) and distal cholangiocarcinoma (DCC) based on the confluence of cystic duct and common hepatic duct.
The utility of 18F-FDG PET/CT in diagnosing primary CCA is associated with the location of the primary lesion[8]. Kim et al[15] found that the sensitivity, specificity, and accuracy of 18F-FDG PET/CT in detecting primary lesions of CCA were 84.0%, 79.3%, and 82.9%, respectively, which were comparable to those of enhanced CT and MRI. Several studies have indicated that 18F-FDG PET/CT exhibits significantly greater diagnostic accuracy for primary lesions of ICC compared to ECC. Albazaz et al[16] found that the overall sensitivity of 18F-FDG PET/CT in the detection of primary CCA was 77%, and the sensitivity of ICC was higher than that of ECC (92%vs 52%). Similar to the above study, Corvera et al[17] found that 18F-FDG PET/CT had a sensitivity of 95% for the detection of ICC, but only 69% for ECC. The discoveries mentioned above may due to the invasive nature and low cell density of ECC.
The 18F-FDG PET/CT has important value in the diagnosis of regional LN metastasis of CCA. A meta-analysis compared the diagnostic value of MRI and 18F-FDG PET/CT for LN metastasis in CCA. The specificity of 18F-FDG PET/CT was significantly higher than that of MRI (0.92 vs 0.52). LN status is an important prognostic factor for the survival of patients diagnosed with CCA, and the identification of LN status has an important impact on treatment management[18]. Overall, 18F-FDG PET/CT demonstrates superior efficacy in assessing LN metastasis in patients with CCA when compared to MRI and MDCT. However, negative results cannot be used as a basis for excluding LN dissection[19].
The occurrence rate of distant metastasis in CCA is substantial, with frequent metastatic locations encompassing the liver, lungs, bones, and brain. The diagnosis and surgical treatment of distant metastatic sites are helpful to improve the tumor-specific survival rate[19]. Kim et al[15] showed that 18F-FDG PET/CT had higher diagnostic accuracy for distant metastases than CT (88.3% vs 78.7%) and provided additional information for lesions that could not be determined on conventional imaging. Different from ECC, 18F-FDG PET/CT has higher diagnostic accuracy for distant metastases of ICC, which may be due to the different tumor metabolic characteristics of distant metastases between ICC and ECC[20]. The 18F-FDG PET/CT is highly sensitive in detecting regional and distant metastatic disease, which helps to avoid unnecessary surgery.
The appearance of false-negative and false-positive outcomes in 18F-FDG PET/CT can be linked to the lack of specificity of FDG as an oncological imaging agent, causing FDG accumulation in particular non-malignant conditions.
ICC
ICC represents an epithelial malignant neoplasm originating from the secondary bile duct and its branches, constituting approximately 20% of all CCA cases, with a steadily rising incidence observed annually[21]. In terms of tumor morphology, ICC was mainly mass type and mixed type (about 90%).
The primary application of 18F-FDG PET/CT in individuals with ICC is to identify potential distant metastatic lesions rather than the primary tumor itself[22]. However, a meta-analysis by Annunziata et al[23] showed that 18F-FDG PET/CT can accurately detect ICC with combined sensitivity and specificity of 95% and 93%, respectively. The typical FDG uptake of ICC was increased in the shape of rosettes, nodules, and masses. CT imaging disclosed the occurrence of hypodense lesions, with or without associated indirect findings, encompassing dilation of the bile duct, existence of bile duct stones, and shrinking of the liver capsule. In addition, it has been suggested that the higher sensitivity of ICC may be related to tumor size, which is usually large[15].
Previous studies have shown that compared with CT and MRI, 18F-FDG PET/CT has better diagnostic value in detecting ICC LN metastasis[24,25]. Park et al's study showed that the sensitivity and specificity of 18F-FDG PET/CT in detecting metastatic LNs were 80% and 92%, respectively[26]. In clinical settings, LN dissection is not typically performed for ICC; however, for patients with positive LNs on 18F-FDG PET/CT, routine LN dissection may be contemplated to enhance staging accuracy and recurrence-free survival rates[22]. Conventional imaging modalities, including CT and MRI, generally identify LN metastasis based on a LN short-axis diameter > 1 cm. However, when the shortest diameter of LNs was < 1 cm, nearly 10% of them still had metastasis. In 18F-FDG PET/CT, LN SUVmax ≥ 2.5 was defined as positive, that is, LN metastasis[26]. FDG uptake value is not affected by LN size. It can provide clearer and visible images under high SUV conditions.
The 18F-FDG PET/CT has clinical significance in the diagnosis of distant metastasis of ICC[27]. The study by Tsurusaki et al[28] found that distant metastasis of ICC may be missed by CT or MRI, while 18F-FDG PET/CT can accurately detect it. A retrospective study included 78 suspected ICC patients who underwent 18F-FDG PET/CT and enhanced CT examination. In the diagnosis of clinical staging, the accuracy, sensitivity, and specificity of 18F-FDG PET/CT diagnosis were 92.31%, 93.54%, and 87.50%, respectively. All of them were significantly higher than those of enhanced CT, providing an important basis for clinical staging. The 18F-FDG PET/CT scanning aids in the detection of distant metastases in ICC, facilitates precise clinical staging, allows patients to undergo treatment strategies divergent from the initial plan, and decreases the incidence of futile surgical interventions.
Some well-differentiated CCA tumors have low intracellular FDG uptake; therefore, false negatives will occur. FDG uptake may be increased in inflamed tissue due to activated macrophages, which can result in false-positive interpretations. All the above conditions will affect the efficiency of 18F-FDG PET/CT in the detection of ICC.
ECC
The incidence of ECC is significantly higher than that of ICC. ECC mainly includes PCC and DCC, among which the incidence of PCC is relatively high, accounting for 60%–70% of the total number of ECCs and 40%–60% of all CCAs[29]. However, there have been few studies on 18F-FDG PET/CT for ECC. Existing research predominantly centers on PCC, with a notable absence of independent studies pertaining to DCC.
In the case of ECC, guidelines and consensus statements advocate the use of 18F-FDG PET/CT to ascertain the presence or absence of tumor metastases, rather than for the diagnosis of the primary extrahepatic tumor itself. A 2018 retrospective analysis confirmed that, compared with enhanced CT, 18F-FDG PET/CT has a higher coincidence rate between the detection of primary lesions of ECC and pathological diagnosis, indicating good diagnostic value for ECC with different properties and differentiation degrees, and its importance for preoperative diagnosis of ECC. Contrary to the above study, Li et al[30] reported that 18F-FDG PET/CT was used for preoperative evaluation of PCC, and the results showed that the sensitivity of the diagnosis of the primary tumor was only 58.8%. This may have been due to the fact that most ECCs are mucinous adenocarcinomas or papillary tumors histologically, and mucin itself does not uptake FDG, so a small number of tumor cells may not be sufficient to generate a detectable PET signal, especially in the surrounding high background liver activity[31]. ECC usually shows a desmoplastic stromal reaction, with tumor cells loosely dispersed within the connective tissue[32].
The 18F-FDG PET/CT has high diagnostic value for metastatic LNs in ECC. A study in 2018 found that the sensitivity, specificity, and accuracy of 18F-FDG PET/CT in the diagnosis of LN metastasis of PCC were 67.9%, 88.0%, and 77.4%, respectively[33]. The research findings mentioned above indicate that 18F-FDG PET/CT is advantageous for the detection of LN metastasis in ECC. However, inflammatory periportal lymphadenopathy can lead to false-positive PET results.
Guidelines and consensus recommend 18F-FDG PET/CT to determine the presence or absence of tumor metastasis. Ruys et al[34] reported that the specificity of 18F-FDG PET/CT in judging distant metastasis of ECC can be as high as 96%. The existence or not of distant metastasis exerts a significant influence on the selection of surgical methods and the prognostic outcome. In a retrospective study in 2022, the sensitivity, specificity, and accuracy of 18F-FDG PET/CT in the diagnosis of distant metastasis of PCC were 47.1%, 97.2%, and 81.1%, respectively. It is suggested that 18F-FDG PET/CT has certain value in identifying unresectable tumors[33]. A limited number of reports suggest that 18F-FDG PET/CT exhibits superior value compared to CT or MRI in the detection of distant metastases in ECC. However, a study in 2020 showed that there was no significant difference in the sensitivity and specificity of 18F-FDG PET/CT, enhanced CT, and MRI in detecting distant metastasis[35].
Several studies have revealed that 18F-FDG PET/CT has importance in the diagnosis of moderately and poorly differentiated non-nodular ECC. Patients with poor differentiation have faster tumor proliferation and higher FDG uptake[36]. Conversely, cancers with poor differentiation typically exhibit a higher density of microvessels. Research findings from studies on lung and renal cancers, among others, indicate a positive correlation between microvessel density and FDG uptake[37]. The 18F-FDG PET/CT demonstrates higher sensitivity in the diagnosis of DCC compared to PCC. This disparity may be explained by the increased metabolic activity of the liver tissue in the proximity of the hilar region.
GBC
GBC refers to malignant tumors occurring in the fundus, body, neck, and cystic duct of the gallbladder. Despite its relative rarity, GBC represents the most prevalent malignant tumor of the biliary system, characterized by its aggressive invasiveness, high lethality, and poor prognostic outcomes[38]. GBCs of epithelial origin are usually adenocarcinoma (95.7%) and occasionally squamous (2.4%) or adenosquamous carcinoma (1.9%); both of which are more aggressive and have a worse prognosis than adenocarcinoma[39]. Presently, the body of literature investigating the application of 18F-FDG PET/CT in GBC is sparse and largely derived from studies with small sample sizes.
The 18F-FDG PET/CT is not usually used for the diagnosis of primary GBC in clinical practice. However, recent studies have revealed the value of 18F-FDG PET/CT in the diagnosis of primary GBC. Adenocarcinoma of the gallbladder usually has FDG uptake, but the degree can vary. A meta-analysis report by Annunziata et al[40], which included a total of 21 studies, showed that the sensitivity and specificity of 18F-FDG PET/CT in the diagnosis of GBC were 87% and 78%, respectively. The authors hold the opinion that 18F-FDG PET/CT, when utilized independently, offers convenience and efficacy; however, it is important to acknowledge the potential for false-negative and false-positive outcomes. In addition, a meta-analysis in 2021 also demonstrated the value of 18F-FDG PET/CT in the diagnosis of GBC, with overall sensitivity and specificity of 96% and 91%, respectively[41].
GBC exhibits a tendency for LN metastasis. Enlarged LNs may result from inflammatory hyperplasia, while LNs of normal size do not preclude the possibility of metastatic involvement. Therefore, 18F-FDG PET/CT is of value in the diagnosis of GBC LN metastasis. A prospective cohort study including 42 patients with suspected gallbladder malignancy demonstrated that the diagnostic sensitivity of 18F-FDG PET/CT for LN metastases was 88.9%. Conventional imaging is difficult for detection of retroperitoneal LN metastasis, which is the most common site of GBC metastasis. Therefore, scholars recommend the use of 18F-FDG PET/CT in GBC patients[42].
The 18F-FDG PET/CT is helpful in detecting distant metastatic disease of GBC and is expected to significantly reduce the number of patients undergoing ineffective surgery[43]. A meta-analysis found that the pooled sensitivity and specificity of 18F-FDG PET/CT for detecting metastatic disease were 95% and 97%, respectively. The 18F-FDG PET/CT changed the treatment of a significant proportion of patients[41]. Some studies have used 18F-FDG PET/CT as an important evaluation method for reoperation in patients with incidental GBC after cholecystectomy.
Local FDG enrichment is suggestive of malignancy but does not distinguish between primary GBC and metastasis from another malignant disease, such as hepatocellular carcinoma (HCC). Furthermore, the pathological progression of benign conditions can influence FDG uptake, leading to false-positive findings.
APPLICATION OF 18F-FDG PET/CT METABOLIC PARAMETERS IN BTC
BTC
18F-FDG PET/CT metabolic parameters can predict the prognosis of patients with BTC: Several studies have proved the value of 18F-FDG PET/CT metabolic parameters in predicting the prognosis of BTC. In each study, the parameters of focus encompassed SUVmax, total lesion glycolysis (TLG), metabolic tumor volume (MTV), as well as the alterations in SUVmax observed during the course of treatment (ΔSUVmax). The primary endpoints were overall survival (OS), disease-free survival (DFS), and progression-free survival (PFS).
In a retrospective study conducted by Ma et al[20], which included 66 patients with BTC who underwent radical surgery, multivariate analysis revealed that SUVmax emerged as an independent prognostic factor influencing DFS and OS. A 2024 meta-analysis confirmed, through pooled data analysis, that an elevated maximum SUVmax was significantly correlated with poor OS and DFS[44]. The study additionally discovered that elevated TLG levels were correlated with reduced OS and DFS. MTV demonstrated a significant association with the risk of mortality and recurrence. The ΔSUVmax emerged as a prognostic indicator for poor OS, yet it failed to demonstrate predictive value for PFS. The above meta-analysis found that the SUVmax of metastatic LNs was not associated with OS. However, a retrospective study in 2022 found that primary tumor and metastatic LN SUVmax ≥ 2.8 was a poor prognostic factor for recurrence-free survival in patients with BTC[45]. Therefore, more studies are needed to evaluate the prognostic value of LN SUVmax.
Several studies have conducted independent investigations into the application of metabolic parameters obtained from 18F-FDG PET/CT for predicting advanced BTC. A retrospective study by Cho et al[46] involving patients with advanced BTC who underwent 18F-FDG PET/CT before palliative chemotherapy showed that higher baseline SUVmax was associated with broader FDG uptake, including more organs (≥ 3) and more lesions (≥ 4). This correlation suggests that tumors with higher SUVmax are more likely to metastasize. The study also revealed that an increase in tumor burden was associated with an increase in SUVmax, which aligns with the previous report indicating a significant correlation between SUVmax and pathological stage[47]. Clinical treatment with gemcitabine combined with cisplatin significantly improved the survival rate of patients. In 2018, Braghiroli et al[48] conducted a study on patients with advanced BTC undergoing first-line chemotherapy with gemcitabine and cisplatin. The patients exhibiting a lower TLG value at baseline or following two cycles of treatment had a prolonged median survival time compared to those with a higher TLG value at baseline. In both studies, patients with higher TLG decline also had longer median survival time. This study suggested that 18F-FDG PET/CT could be used as an effective means to predict the prognosis of patients with advanced BTC treated with first-line chemotherapy with gemcitabine and cisplatin.
CCA
18F-FDG PET/CT metabolic parameters can predict the prognosis of patients with CCA: The 18F-FDG PET/CT metabolic parameters have been widely used in predicting the prognosis of patients with CCA. Sabaté-Llobera et al[49] conducted a retrospective analysis of 60 patients with CCA who had not received prior treatment and discovered that, in the multivariate analysis, advanced disease stage, increased levels of carcinoembryonic antigen (CEA), were significantly linked to reduced OS. Metabolic parameters and tumor marker levels were correlated with tumor burden and exerted an influence on prognosis. The study also found significant metabolic differences between ICC and ECC, and ICC showed a significantly higher affinity for FDG, mainly for two reasons: (1) ICC and ECC have different cellular origins. Reprogramming mechanisms can induce cellular plasticity, which can promote ICC to originate from any hepatocyte[50]. In contrast, ECC originates from bile duct epithelium and peribiliary glands[51]. Different origins may imply differences at the molecular level, which can lead to different FDG affinity and metabolic patterns; and (2) Different macro-growth patterns. Mass type is the most common type of CCA, especially in ICC. The sensitivity of PET in detecting mass-type CCA is higher than for other types. However, in this study, although the metabolic behavior of the two was different, it did not affect the prognosis.
Apart from the aforementioned conventional metabolic indices, novel metabolic parameters have emerged in the last 2 years for predicting the outcome of patients with CCA. In 18F-FDG PET/CT, FDG uptake by the reticuloendothelial system has been associated with systemic inflammatory responses (SIRs) to cancer cells in patients with a variety of malignancies[52]. In 2024, scholars from Korea investigated the prognostic significance of FDG uptake in predicting PFS and OS[53]. Their findings revealed that the bone marrow-to-aorta uptake ratio (BAR) exhibited significant correlations with tumor stage and serum inflammatory biomarkers. In multivariate survival analysis, BAR is an independent predictor of PFS and OS. This study highlights that BAR is a promising independent predictor with the potential for personalized prognosis and treatment strategies. Previous studies have shown that visceral adiposity is associated with poor prognosis in patients with CCA. In 2024, Lee et al's group also investigated the effect of adipose tissue imaging parameters, which reflect quantitative and qualitative characteristics of subcutaneous adipose tissue and visceral adipose tissue (VAT), on the prognosis of 94 resected patients with CCA[54]. Multivariate survival analysis demonstrated that the visceral-to-subcutaneous adipose tissue area ratio (VSR) and the average FDG uptake in VAT served as predictors of RFS. This study suggests that FDG uptake in VSR and VAT may be promising prognostic indicators for predicting RFS in patients with CCA.
ICC
18F-FDG PET/CT metabolic parameters can predict the prognosis of patients with ICC: The 18F-FDG PET/CT metabolic parameters have important effects on predicting the prognosis of ICC patients. A retrospective analysis in 2022, involving 291 consecutive ICC patients who underwent 18F-FDG PET/CT[55], suggested that patients with high trap-neuter-release (TNR) had significantly shorter OS and RFS, as did those with high tumor SUVmax. In multivariate analysis, the OS and RFS were significantly shorter in patients with high TNR, as were those with high tumor SUVmax. The TNR, rather than the maximum SUVmax of the tumor, exerted a significant influence on OS and RFS, implying that TNR serves as a superior prognostic marker compared to tumor SUVmax. Seo et al[56] evaluated the prognostic relationship between FDG uptake and SIR markers in ICC patients, and selected neutrophil to lymphocyte ratio (NLR) as SIR markers. Multivariate cox regression analysis found that, high tumor SUVmax (≥ 8) and high NLR (≥ 5) were independent predictors of poor OS and DFS. This researcher suggested that tumor SUVmax and SIR markers can effectively predict the prognosis of ICC patients[56]. Body mass index (BMI) is known to be associated with poor prognosis in several cancers[57]. Yugawa et al[58] investigated the association between BMI and patient prognosis, discovering that an elevated BMI constituted an independent prognostic factor for unfavorable outcomes and heightened recurrence risk in patients with ICC following radical hepatectomy. Importantly, BMI was highly correlated with FDG accumulation on 18F-FDG PET/CT imaging. These observations suggest that obesity is a risk factor for cancer progression associated with altered metabolic activity and immune status in ICC.
18F-FDG PET/CT metabolic parameters can predict gene expression in patients with ICC: Recent efforts have concentrated on utilizing 18F-FDG PET/CT metabolic parameters for predicting gene expression patterns. In 2019, Ahn et al[59] carried out RNA sequencing on 22 patients with ICC who had undergone 18F-FDG PET/CT scanning prior to surgery, to assess gene expression profiles associated with FDG uptake patterns. The subsequent analysis indicated that pathways linked to the cell cycle, mitosis, hypoxia, inflammatory response, and metabolism were over-represented in patients with elevated SUVmax values. They also found that the FDG uptake profile of ICC patients can predict genomic features of gene expression and pathways, and that cell cycle, metabolic and hypoxic pathways are enriched in patients with high FDG uptake, which may guide more rational targeted therapy. Ikeno et al[60] demonstrated an association between elevated MTV and TLG values and the presence of KRAS mutations within tumors. MTV level of ≥ 38 was found to predict a reduced 5-year survival rate, with an overall accuracy of 68%, specificity of 67.9%, and sensitivity of 77.8% in forecasting KRAS mutations. The study could not only provide noninvasive assessment of genetic composition, but also provide prognostic assessment of the disease, as patients with tumors with KRAS mutations are known to have poor prognosis. A retrospective study by Zhang et al[61] also found that SUVmax and TNR of moderately and poorly differentiated ICC were significantly correlated with tumor differentiation, size, and Ki67 expression.
ECC
18F-FDG PET/CT metabolic parameters can predict the prognosis of patients with ECC: Some studies have explored the application of metabolic parameters derived from 18F-FDG PET/CT in predicting the prognostic outcomes of patients with ECC. In 2023, a retrospective study enrolled 86 patients with ECC, and revealed two points. First, the postoperative recurrence rate of the high SUV group (SUVmax ≥ 4.9) was higher than that of the low SUV group, and the expression rate of Glut1 and Ki-67 was higher in the high SUV group. Secondly, by immunostaining of resected specimens, the high SUV group was correlated with the histological malignancy of the tumor[62]. In the study by Lee et al[63], multivariate analysis showed that TLG was an independent factor affecting the prognosis of ECC patients. SUVmax was not an independent prognostic factor. Previous studies have shown that SUVmax is associated with tumor aggressiveness and survival of patients with various cancer types. However, it has limited value for accurately showing the metabolic load of the whole tumor and not the metabolic activity of the whole tumor; therefore, some reports have suggested that tumor volume parameters measured by 18F-FDG PET/CT, such as TLG and MTV, are more accurate prognostic factors than SUVmax in patients with various malignancies. Yi et al[64] found that only low SUVmax (≤ 2.7) was an independent prognostic factor associated with poor OS in DCC in multivariate analysis, which contradicted the results of previous studies. Therefore, further extensive research is needed to investigate the relationship between tumor FDG avidity, growth patterns, and the prognostic value of metabolic parameters obtained from 18F-FDG PET/CT in DCC.
GBC
18F-FDG PET/CT metabolic parameters can predict the prognosis of patients with GBC: Previous research has indicated that 18F-FDG PET/CT metabolic parameters possess some predictive value for assessing the prognosis of patients with GBC. Chun et al[65] retrospectively analyzed 83 patients with locally advanced or metastatic GBC, all of whom underwent 18F-FDG PET/CT scanning at the time of initial diagnosis. In univariate analysis, pathological differentiation, performance status, C-reactive protein level, highest SUV in metastases (SUVmt max) and the sum of MTV of primary tumor and metastases (MTVtotal) were significantly associated with OS. In multivariate analysis, MTVtotal remained significant for predicting OS. The authors emphasized that 8F-FDG PET/CT parameters based on total tumor burden, such as MTVtotal, are of value in identifying patients with poor prognosis in locally advanced and metastatic GBC. A retrospective study in 2022 included 54 patients with GBC[66]. In multivariate analysis, only clinical stage and MTV were significant, and MTV had the highest odds ratio. Pearson correlation coefficient showed that PFS was moderately negatively correlated with clinical stage and MTV. The present study revealed that clinical stage and MTV are the most reliable parameters for forecasting recurrence and disease progression in adenocarcinoma of the gallbladder. Based on clinical stage, MTV would represent a strong prognostic predictor.
18F-FDG PET/CT metabolic parameters are helpful for the differential diagnosis of benign and malignant gallbladder diseases: Currently, surgical resection remains the sole curative option for GBC, with R0 resection being a crucial determinant of prognosis. In clinical practice, the majority of cases is diagnosed through histopathological examination of gallbladder specimens following cholecystectomy[67]. For patients with cholecystitis, treatment, including supportive care and surgery, depends mainly on the type of cholecystitis. Therefore, preoperative judgment of GBC and cholecystitis helps patients to choose the correct treatment[68]. In 2023, Li et al[68] demonstrated the utility of a multi-parameter model, which incorporates 18F-FDG PET/CT metabolic parameters and clinical indicators, for differentiating between nonmetastatic GBC and cholecystitis. The study comprised 88 patients diagnosed with GBC and 34 with cholecystitis, aimed at assessing PET metabolic parameters, including SUVmax, SUVmean, SUVpeak, MTV, TLG, and tumor/normal liver SUV ratio (SUVR). The receiver operating characteristic curve analysis revealed that SUVR exhibited the highest diagnostic accuracy among all metabolic parameters, with a sensitivity of 0.341, specificity of 0.971, positive predictive value of 0.968, and negative predictive value of 0.363. The area under the curve (AUC) of the combined diagnosis model of gallstones, fever, CEA > 5 ng/mL, and SUVR was 0.899, and the diagnostic efficacy of the model was significantly higher than that of SUVR. This study suggests that a diagnostic model incorporating PET metabolic parameters, specifically SUVR, along with clinical variables, demonstrates high diagnostic efficacy in differentiating non-metastatic GBC from cholecystitis, thereby facilitating appropriate treatment decisions for patients.
APPLICATION OF 18F-FDG PET/CT RADIOMICS IN BTC
18F-FDG PET/CT radiomics
Radiomics is a high-throughput, noninvasive medical imaging technique commonly used in recent years. This innovative technology uses computer algorithms to define the region of interest (ROI) in medical images and extract pertinent image features, enabling a quantitative and objective characterization of lesions. Radiomics can be divided into five steps: (1) Data acquisition; (2) ROI analysis; (3) Feature extraction; (4) Statistical analysis, model establishment and application; and (5) Finally applied to the auxiliary diagnosis of diseases[69]. The 18F-FDG PET/CT combines the high-resolution anatomical imaging capabilities of CT with the high specificity of functional metabolic imaging provided by PET. 18F-FDG PET/CT radiomics can collect more complete data from PET images, CT images, and fusion images, and quantify and quantitatively analyze them to obtain a series of molecular imaging level information such as pathological and physiological changes and gene expression status of lesions, so as to describe lesions more completely and provide more accurate and reliable diagnosis and prognosis information. It provides a reference for the personalized treatment of patients.
Application of 18F-FDG PET/CT radiomics in BTC
Recent studies have shown that 18F-FDG PET/CT radiomics has good clinical application value and development prospects. However, in terms of BTC, 18F-FDG PET/CT radiomics studies have focused on the subtype of ICC, which has been helpful in disease diagnosis, predicting pathological information, and predicting prognosis. Existing studies present certain constraints. Predominantly, they use a limited sample size and a retrospective, single-center study design, potentially resulting in model instability and subsequently impacting the precision of feature values and predictive outcomes. In the future, larger scale studies are still needed to verify.
18F-FDG PET/CT radiomics was used to predict the pathological information of ICC: ICC is an aggressive tumor with increasing incidence and few reliable biomarkers[70]. Tumor malignancy is primarily determined by its grade and the existence of microvascular invasion (MVI), which can only be definitively ascertained through the examination of surgical specimens. To date, ICC treatment planning has relied on morphological criteria, such as tumor shape and size, and carbohydrate antigen (CA) 19-9 values, which are poor proxies for tumor biology[71]. In 2022, the study of Fiz et al[72] included 74 patients who underwent hepatectomy for mass-like ICC. On the PET images, ICC was manually segmented (tumor-VOI), and a 5-mm margin around the tumor (margin-VOI) was semi-automatically generated, and texture analysis was performed using LifeX software. The clinical + radiomics model considering tumor grade, MVI, and tumor-VOI simultaneously had better grading performance than the clinical model (AUC = 0.78 vs AUC = 0.72; MVI: 0.87 vs 0.78), which achieved noninferior performance to the postoperative model. Integration of margin-VOI into the radiomics model led to an improved prediction of tumor grade, although its effect on predicting MVI was not significant. Simultaneously, it was observed that tumors exhibiting linear patterns of low uptake were associated with less-aggressive characteristics (G1-2 or MVI-negative). This study revealed that 18F-FDG PET/CT texture analysis was helpful for preoperative assessment of tumor aggressiveness in patients with mass-like ICC and obtained preoperative prediction that can usually only be achieved after resection with pathological data. As an innovative noninvasive biomarker, PET radiomics is helpful for the accurate assessment of ICC patients.
MVI is an important risk factor for early recurrence of ICC. Jiang et al[73] included 51 patients with suspected MVI and ICC (MVI positive:negative = 31:20). The prediction model used a supervised machine learning classifier. The AUC of the MVI prediction model based on two PET features and CA19-9 was 0.90 (accuracy, sensitivity, and specificity were 0.77, 0.75, and 0.80, respectively). The 18F-FDG PET/CT radiomics features have potential value in the prediction of MVI in patients with ICC, and this prediction model can provide noninvasive biomarkers for early prediction and comprehensive quantification of ICC.
18F-FDG PET/CT radiomics predicts the prognosis of ICC: As previously reported for CT and MRI[74], radiomics has improved the predictive/prognostic performance for clinical variables. However, only a few studies have revealed the prognostic value of 18F-FDG PET/CT radiomics in ICC patients. In 2024, Kwon et al[75] developed a radiomics model utilizing machine learning techniques, leveraging 18F-FDG PET/CT data, to forecast the RFS and OS of patients with ICC. Radiomics features such as first order, shape, and gray scale were extracted from the scanned images of 52 patients. Unsupervised hierarchical clustering was used to analyze the radiomics features [group 1 (n = 27); group 2 (n = 23); group 3 (n = 2)]. Multivariate analysis showed that the PET radiomics group was an independent prognostic factor for RFS and OS. The findings of this research suggest that 18F-FDG PET/CT radiomics holds promise for preoperative prognosis prediction in ICC patients, as it minimizes the inaccuracies associated with biopsy samples through its noninvasive and reliable imaging approach. In addition, radiomics can be used as a tool to help detect the expression of specific genes in biopsy material or to determine the optimal location of the biopsy site, and to provide valuable complementary information for genomic analysis of ICC patients. A retrospective study in 2022 involving 51 ICC patients demonstrated that the tumor-VOI and margin-VOI models, incorporating preoperative clinical data along with radiomic features, outperformed clinical models used in isolation. These combined models were also shown to be on a par with models that included pathological and postoperative information in predicting OS and PFS. The prognostic value of texture features extracted from 18F-FDG PET/CT in mass-type ICC has been demonstrated[72].
Role of 18F-FDG PET/CT radiomics in the differential diagnosis of ICC: HCC and ICC are two common subtypes of primary liver cancer with different prognosis[76]. Therefore, preoperative prediction of HCC and ICC classification is essential to provide accurate information and reasonable treatment. A retrospective study in 2022 included 127 patients with HCC (n = 78) or ICC (n = 51). The AUC of the radiomics prediction model composed of two PET and one CT features in the classification of HCC and ICC in the test cohort was 0.86 (accuracy, sensitivity, and specificity were 0.82, 0.78, and 0.88, respectively). The study also found that factors related to tumor intensity and texture were the most important components in predicting histological classification, which is partially consistent with the findings of Wu et al[77]. The 18F-FDG PET/CT radiomics offers image-based molecular features and insights into intratumoral heterogeneity, showing potential as a valuable diagnostic tool for the histological subclassification of primary liver cancer[78]. In 2024, Diao and Jiang[79] proposed a new tumor subtype classification method based on 18F-FDG PET/CT radiomics and deep learning (RA-DL) for small data sets, and carried out verification experiments on liver cancer, lung cancer, and lymphoma data sets. In the binary classification task of liver tumors (HCC and ICC), the AUCs were 0.84 and 0.86, respectively. This study fused depth features and radiomics features through RA-DL , which can effectively capture ROI. This method can accurately classify tumor subtypes and help to formulate effective treatment plans.
APPLICATION OF PET/MRI IN BTC
The main limitations of PET imaging arise from its low spatial resolution and inability to clearly define anatomical structures of tissues. MRI has advantages over CT in soft tissue imaging. PET/MRI combines physiological and molecular data from PET with functional and anatomical information from MRI, offering advantages in the diagnosis of BTC. Compared with MRI alone, PET/MRI imaging provides patients with necessary functional information that can improve diagnostic accuracy, especially in assessing tumor activity and response to therapy[80].
CCA has a poor prognosis, and surgical resection is the preferred treatment. Accurate staging to determine resectability at diagnosis is essential to guide treatment decisions. PET/MRI has been shown to improve the overall diagnostic accuracy of CCA. A retrospective analysis by Yoo et al[81] showed that PET/MRI changed the clinical management of 30% of patients with newly diagnosed ICC compared with conventional imaging, which could better assess and delineate the extent of intrahepatic lesions, and detect and exclude distant metastases[81]. In 2023, Pang et al[82] investigated the utility of PET/MRI for preoperative staging of HCC and compared it to 18F-FDG PET/CT. PET/MRI improved the accuracy of the Bismuth–Corlette classification and increased the precision of LN staging by 17.3% compared to 18F-FDG PET/CT. In addition, this study showed that PET/MRI and 18F-FDG PET/CT have similar detection efficiency for distant metastasis. Although data on PET/MRI for assessing response in CCA are limited, case reports and institutional experience suggest that PET/MRI is superior to 18F-FDG PET/CT, MRI, and CT alone, especially in patients with elevated tumor markers and negative imaging[81]. Compared to 18F-FDG PET/CT, PET/MRI demonstrated enhancement in evaluating local recurrence. The enhancement can be attributed to the multiparametric information provided by MRI, coupled with the superior soft tissue contrast of MRI.
GBC is the most common malignant tumor of the biliary system. Increased FDG uptake is helpful in differentiating benign from malignant gallbladder wall thickening as shown by ultrasound, CT, or MRI[83]. However, the main application of FDG-PET in GBC is to identify distant metastasis and recurrence. Typical imaging methods for staging GBC include ultrasound, CT, and MRI/magnetic resonance cholangiopancreatography. A significant proportion of GBC is incidentally detected during cholecystectomy. MRI is crucial for evaluating the anatomical structure of the biliary tract and assessing local invasion of the liver parenchyma and LN metastases in GBC. Theoretically, PET/MRI can provide the advantage of MRI in local staging and the advantage of PET in metastatic disease staging, but the role of PET/MRI in GBC staging has yet to be confirmed by studies[84].
