Patterns of presentation and diagnostic delays in biliary tract cancers: a real-world European study Authors: A Martirena, L Cavka, V Guarini, A Lamarca Published: 2026 ABSTRACT AbstractBackgroundBiliary tract cancers (BTCs) often present late due to nonspecific symptoms and a lack of effective population screening. We characterised diagnostic routes, symptom-to-diagnosis intervals, and their relationship with stage and outcomes across three European centres.Materials and methodsWe conducted a retrospective observational study of consecutive patients diagnosed with BTC assessed by medical oncology from 1 June 2022 to 31 January 2024 in Spain, Slovenia, and Italy. Demographics, presenting symptoms, route of diagnosis, investigations, treatments, and outcomes were extracted from electronic medical records. Survival was assessed using Kaplan–Meier analysis and Cox regression. The study was approved by local ethics committees.ResultsEighty-six patients were included [Spain 59 (68.6%), Slovenia 16 (18.6%), and Italy 11 (12.8%)]; the median age was 71 years (range 40-96); and 65.1% were male. Tumour types included intrahepatic cholangiocarcinoma (41.9%), distal cholangiocarcinoma (23.3%), gallbladder (12.8%), ampullary (11.6%), and perihilar cholangiocarcinoma (9.3%). Presentation routes included: primary care (46.5%), the emergency department (45.3%), and incidental (8.1%). The median time from symptom onset to diagnosis was 2.56 months [95% confidence interval (CI) 1.78-3.09], which was longer with primary care (versus emergency department) (4.13 months versus 1.37 months, P = 0.14). According to symptoms, delays were longest for weight loss (3.47 months) and asthenia/anorexia (3.28 months), and shortest for jaundice (1.36 months). Country-level differences in symptom patterns were observed, along with differences in presentation routes. At diagnosis, 55.8% had metastatic disease and 16.3% were diagnosed based on cytology only. Diagnostic work-up varied by country (first imaging ultrasound (Slovenia) versus magnetic resonance imaging (Italy), P < 0.001). Multidisciplinary tumour board discussion occurred in 77.9% of patients. Treatment included surgery with curative intent in 43% of patients (of whom 56.8% received adjuvant chemotherapy) and first-line palliative chemotherapy in 61.6% (of whom 67.6% received subsequent lines). No active anticancer therapy was administered in 12.8% of patients. After a median follow-up of 13.47 months, the estimated median overall survival (OS) was 22.28 months (95% CI 17.42-34.51). In multivariable analysis, curative surgery [hazard ratio (HR) 0.19, 95% CI 0.04-0.86, P = 0.032] and first-line palliative systemic therapy (HR 0.15, 95% CI 0.05-0.44, P < 0.001) were associated with longer OS, whereas male sex (HR 3.69, 95% CI 1.17-11.67, P = 0.026) and jaundice at diagnosis (HR 2.48, 95% CI 1.11-5.54, P = 0.027) were associated with shorter OS. Neither route of diagnosis nor diagnostic delay was associated with OS, although this analysis may have been underpowered.ConclusionBTC continues to be diagnosed at advanced stages and is frequently associated with diagnostic delays, often presenting with nonspecific symptoms. Strengthening referral pathways and improving early recognition of warning signs, particularly in primary care, may help optimise diagnostic processes and should be further evaluated in future studies.Key words: biliary tract cancer, diagnostic delay, early symptoms, real-world study, route of diagnosis, survival outcomes Introduction Biliary tract cancers (BTCs) are typically characterised by two key features: rarity and extreme aggressiveness.1 Given their tendency for early dissemination and poor prognosis, it is not surprising that only 20%-35% of patients with BTC are eligible for curative-intent treatment at diagnosis,2,3 with wide geographic disparities across Europe and around the world, driven mainly by differences in access to specialised centres, diagnostic resources, and multidisciplinary care.4,5 Late diagnosis remains one of the most critical challenges in BTC. The rarity of the disease, together with its heterogeneous and often nonspecific symptoms, which are easily mistaken with more common conditions, frequently contributes to delayed diagnosis.4, 5, 6, 7 As a result, many patients present with advanced disease when curative options are no longer feasible. A recent UK cohort study8 found that more than half of patients diagnosed with cholangiocarcinoma received no cancer-specific treatment at all (neither potentially curative surgery, systemic therapy, nor biliary stent insertion). Moreover, inadequate biliary drainage and biliary tract infections further worsen performance status, often rendering systemic treatment unfeasible.9 Early diagnosis, timely referral, and adequate supportive care are therefore critical. Screening programmes have proven effective in the early detection of many cancers; however, for BTC [outside the setting of primary sclerosing cholangitis (PSC)], no such initiatives have been widely adopted. In patients with PSC, who have a markedly increased risk of developing cholangiocarcinoma, the latest European Association for the Study of the Liver Clinical Practice Guidelines recommend annual surveillance with magnetic resonance imaging (MRI) or Magnetic Resonance Cholangiopancreatography, with or without serum CA19-9 measurement, to enable early detection of malignancy. However, the effectiveness of this approach in reducing mortality remains uncertain.10 In addition, it should be noted that in Western countries, BTC is predominantly sporadic and often lacks identifiable risk factors. A few conditions are able to increase the risk of developing BTCs, such as biliary lithiasis, liver cirrhosis, and various inflammatory and metabolic liver diseases.9 However, the value of screening remains highly debated.7,11 Novel noninvasive methods such as blood-based microRNA profiling show promise but remain unvalidated for clinical use.12 Based on the lack of clear supportive evidence, to date, no health authorities currently endorse population-wide screening for the non-PSC population who may be at risk of developing BTC. In the absence of reliable screening approaches, greater awareness of early symptoms is essential. Approximately half of BTCs are still diagnosed in the emergency department due to acutely worsening symptoms, which may be associated with poorer outcomes.13,14 It has also been shown that patients with BTC (and those with pancreatic cancer) experience mild but continuous and bothersome symptoms even a year before diagnosis, such as fatigue, weight loss, or gastrointestinal complaints, and they frequently have abnormal glucose or bilirubin levels that were not investigated further.6 These findings suggest that a combination of better awareness on warning symptoms and early investigations could support earlier referral. Recent advances in systemic therapy, including immune checkpoint inhibitors and molecularly targeted agents, have expanded treatment options and improved outcomes in selected patients. Nevertheless, the real-world impact of these therapies remains constrained by late diagnosis, disease burden at presentation, and inequalities in access to specialist care and biomarker testing.15 Improving early detection pathways is therefore crucial to translate therapeutic progress into meaningful survival gains. In summary, early and accurate diagnosis of BTC is a huge unmet need. While earlier diagnosis could potentially increase the number of resectable patients, thereby increasing the rate of cure, a timelier diagnosis in patients with advanced disease could also increase the percentage of patients fit for palliative therapies. Understanding patterns of symptom presentation, measuring diagnostic intervals, and identifying factors contributing to delays are essential steps to understand current gaps in care and ultimately improve outcomes.4, 5, 6 Understanding these patterns will also help design future studies exploring early-diagnosis interventions to improve current diagnostic pathways. In this study, we aimed to describe the clinical presentation and diagnostic pathways of patients with BTC to identify unmet needs and opportunities for earlier diagnosis and intervention. Materials and methods This retrospective, observational, multicentre study included three centres in Spain, Italy, and Slovenia. The study aimed to analyse time intervals from symptom onset to histological confirmation, identify factors associated with diagnostic delay, and assess their impact on stage at diagnosis and eligibility for curative treatment. The goal was to inform interventions that could enable earlier diagnosis and improve outcomes for patients with BTC. The study protocol was approved by the ethics committee of each participating centre. Consecutive patients evaluated in the medical oncology departments between 1 June 2022 and 31 January 2024 were included. All types of biliary tract tumours [intrahepatic cholangiocarcinoma (iCCA), perihilar cholangiocarcinoma (pCCA), distal cholangiocarcinoma (dCCA), gallbladder cancer, and ampullary carcinoma] were eligible, provided that a histological or cytological diagnosis was available, as pathological confirmation was an essential inclusion criterion. Patients lacking the minimum clinical data required for analysis, such as an absence of confirmed diagnosis or incomplete information regarding presentation route and clinical characteristics, were excluded. Data were retrospectively collected from medical records, including demographic information, initial symptoms, route of presentation (emergency department, primary care, or incidental diagnosis), tumour stage at diagnosis, treatment intent, and survival outcomes. Time intervals between symptom onset and diagnosis were also assessed. Presenting symptoms were recorded at the time of first clinical evaluation. When multiple symptoms were present, the predominant symptom was defined as the symptom that prompted the diagnostic work-up, as documented in the clinical records. The date of symptom onset was defined as the earliest reported symptom potentially attributable to the disease, based on patient-reported past medical history documented in the medical records. Therefore, time to diagnosis was calculated from the first symptom onset, regardless of the predominant symptom leading to diagnostic evaluation. Given the retrospective design, this variable is subject to recall and documentation bias. Time intervals between symptom onset and diagnosis were also assessed. The primary objective was to describe the presentation pattern of BTCs, including route of diagnosis, presenting symptoms, and time from symptom onset to diagnosis. Secondary objectives included evaluating diagnostic pathways and procedures, treatment approaches at diagnosis, overall survival (OS), and progression-free survival. Descriptive statistics were used to summarise the collected variables. Continuous variables were expressed as the median with minimum and maximum values, and categorical variables as absolute frequencies and percentages. The chi-square test, Fisher’s exact test, Student’s t-test, and analysis of variance were used for exploratory comparisons as appropriate. OS was defined as the time from first diagnosis to death (patients alive at the time of analysis were censored at the time of last visit). Survival analysis was conducted using the Kaplan–Meier method and univariate and multivariable Cox regression. A threshold of P < 0.05 (two-tailed) was considered significant. Results Patient population A total of 86 eligible patients with biliary tract tumours were included. The number of patients per centre was 59 (68.6%) in Spain, 16 (18.6%) in Slovenia, and 11 (12.8%) in Italy. Patient characteristics are summarised in Table 1. The median age was 71 years (range 40-96), and 65.1% of patients were male. Regarding tumour type, iCCA was the most frequent (36 patients, 41.9%), followed by dCCA (20 patients, 23.3%), gallbladder carcinoma (11 patients, 12.8%), ampullary carcinoma (10 patients, 11.6%), and pCCA (8 patients, 9.3%). Most patients did not have a significant past medical history (79.1%) or family history (75.6%) of cancer. Baseline characteristics were generally similar across countries, although differences in sex distribution were observed, with a higher proportion of male patients in the Italian (90.9%) and Slovenian (81.3%) cohorts compared with the Spanish cohort (55.9%) (P = 0.031). Table 1Baseline demographic and clinical characteristics of the study population by countryVariableAll (N = 86)Spain (N = 59)Slovenia (N = 16)Italy (N = 11)P valueSex0.031 Male56 (65.1)33 (55.9)13 (81.3)10 (90.9) Female30 (34.9)26 (44.1)3 (18.8)1 (9.1)Age, median (min-max)71 (40-96)70 (40-90)73 (48-96)72 (43-83)0.506First-degree family history0.065 No65 (75.6)46 (78)9 (56.3)10 (90.9) Yes—malignant16 (18.6)8 (13.6)7 (43.8)1 (9.1) Yes—benign5 (5.8)5 (8.5)00Additional past medical history0.247 No68 (79.1)44 (74.6)13 (81.3)11 (100) Yes—malignant10 (11.6)7 (11.9)3 (18.8)0 Yes—benign8 (9.3)8 (13.6)00Type of BTC0.058 iCCA36 (41.9)26 (44.1)2 (12.5)8 (72.7) hCCA8 (9.3)6 (10.2)2 (12.5)0 dCCA20 (23.2)11 (18.6)7 (43.8)2 (18.2) GBC11 (12.8)10 (16.9)1 (6.3)0 AMP10 (11.6)6 (10.2)3 (18.8)1 (9.1) No data1 (1.2)01 (6.3)0Values are expressed as n (%) unless otherwise specified. P values were calculated using the chi-square test or Fisher’s exact test as appropriate.AMP, ampullary cancer; BTC, biliary tract cancer; dCCA, distal cholangiocarcinoma; GBC, gallbladder cancer; HCCA, hilar cholangiocarcinoma; iCCA, intrahepatic cholangiocarcinoma; max, maximum; min, minimum. Baseline demographic and clinical characteristics of the study population by country Values are expressed as n (%) unless otherwise specified. P values were calculated using the chi-square test or Fisher’s exact test as appropriate. AMP, ampullary cancer; BTC, biliary tract cancer; dCCA, distal cholangiocarcinoma; GBC, gallbladder cancer; HCCA, hilar cholangiocarcinoma; iCCA, intrahepatic cholangiocarcinoma; max, maximum; min, minimum. Route of presentation and symptoms at diagnosis The route of presentation was primary care in 46.5% of patients (40/86), the emergency department in 45.3% (39/86), and incidental diagnosis in 8.1% (7/86). The distribution was similar in Spain and Slovenia. In Italy, a higher proportion of patients were diagnosed in primary care, although this difference was not statistically significant (P = 0.754) (Figure 1A). Route of presentation and presenting symptoms in patients with biliary tract cancer. (A) Distribution of routes of presentation (primary care, emergency department, and incidental diagnosis) overall and by country (Spain, Slovenia, and Italy). Percentages refer to the proportion of patients in each country. No statistically significant differences were observed among the countries. (B) Frequency of presenting symptoms at diagnosis. Bar length represents the proportion of patients with each symptom (percentage). Colours indicate differences by route of diagnosis: red, more frequent in the emergency department; blue, more frequent in primary care; grey, similar frequencies between routes. Jaundice was more frequent in patients presenting to the emergency department, whereas increased transaminase and asthenia/anorexia were more frequent in patients referred to primary care. No significant differences were observed for the remaining symptoms. The most frequent presenting symptoms in patients were jaundice (39.5%), pain (38.4%), increased transaminases (36%), weight loss (32.6%), asthenia and/or anorexia (23.3%), dyspepsia (9.3%), and diarrhoea (5.8%). Other symptoms were reported in 19.8% of patients. Differences in symptom distribution across countries are summarised in Table 2. For example, lower rates of pain and higher rates of asthenia/anorexia were observed in Italy (P = 0.041 and P < 0.001, respectively), while Slovenia showed a higher proportion of jaundice at presentation (P = 0.001). Table 2Presenting symptoms of patients with biliary tract cancer by country and route of diagnosisSymptomAll (N = 86)Spain (N = 59)Slovenia (N = 16)Italy (N = 11)P value (country)Primary care (N = 40)Emergency department (N = 39)P value (route of diagnosis)Pain33 (38.4)23 (39)9 (56.3)1 (9.1)0.04116 (40.0)17 (43.6)0.821Jaundice34 (39.5)18 (30.5)13 (81.3)3 (27.3)0.0019 (22.5)25 (64.1)<0.001Weight loss28 (32.6)12 (20.3)9 (56.3)7 (63.6)0.00214 (35.0)14 (35.9)0.562Increased transaminases31 (36)17 (28.8)13 (81.3)1 (9.1)<0.00122 (55.0)9 (23.1)0.005Dyspepsia8 (9.3)3 (5.1)4 (25)1 (9.1)0.0445 (12.5)3 (7.7)0.712Diarrhoea5 (5.8)3 (5.1)2 (12.5)00.3393 (7.5)2 (5.1)1.000Asthenia and/or Anorexia20 (23.3)5 (8.5)6 (37.5)10 (90.9)<0.00112 (30.0)8 (20.5)0.439Other17 (19.8)11 (18.6)4 (25)2 (18.2)0.9137 (17.5)10 (25.6)0.422Values are expressed as n (%). P values were calculated using the chi-square test or Fisher’s exact test (two-tailed) as appropriate. Presenting symptoms of patients with biliary tract cancer by country and route of diagnosis Values are expressed as n (%). P values were calculated using the chi-square test or Fisher’s exact test (two-tailed) as appropriate. When symptoms were explored by route of diagnosis across all countries together (Table 2), jaundice was more frequent among patients presenting to the emergency department compared to primary care (25 (64.1%) versus 9 (22.5%) patients, P < 0.001). In contrast, increased transaminases were more frequent in patients referred to primary care (22 patients compared with 9 patients, P = 0.005). Asthenia and/or anorexia were also more frequent in patients diagnosed in primary care, although this difference was not statistically significant. No significant differences were observed for weight loss, dyspepsia, diarrhoea, and other nonspecific symptoms (all P > 0.05) (Figure 1B). Time to diagnosis The median time from symptom onset to diagnosis was 2.56 months [95% confidence interval (CI) 1.78-3.09]. By route of presentation, the median time to diagnosis was longer in patients presenting to primary care (versus emergency department) (4.13 months versus 1.37 months, P = 0.14). When analysed by country, the median time to diagnosis was consistently longer in patients diagnosed in primary care in each cohort, although differences were not statistically significant. No differences in time to diagnosis were observed by sex (2.56 months for men versus 2.61 months for women, P = 0.944) or age group (2.04 months for patients <65 years versus 2.91 months for those ≥65 years, P = 0.57) (Table 3). Table 3Time from symptom onset to diagnosis according to route of presentation and ageTime (months) to diagnosis, median (95% CI)P valueAll (N = 79, excludes incidental diagnosis)2.56 (1.78-3.09)PC (N = 40)4.13 (2.96-5.19)0.140ED (N = 39)1.37 (0.66-2.30)Spain (N = 54)2.15 (1.18-3.29)0.413 PC (N = 26)4.59 (2.96-8.97)0.145 ED (N = 28)1.12 (0.39-1.68)Slovenia (N = 15)2.56 (1.41-3.02) PC (N = 7)2.56 (0.76-5.16)0.858 ED (N = 8)2.59 (0.59-10.61)Italy (N = 10)3.38 (1.48-6.70) PC (N = 7)4.26 (1.25-11.46)0.072 ED (N = 3)2.16 (1.48-2.99)Age <65 years (N = 27)2.04 (1.08-5.19)0.577Age ≥65 years (N = 52)2.91 (1.77-3.28)Male (N = 53)2.56 (1.68-3.28)0.944Female (N = 26)2.61 (1.68-3.28)P values were calculated using nonparametric tests, as appropriate.CI, confidence interval; ED, emergency department; PC, primary care. Time from symptom onset to diagnosis according to route of presentation and age P values were calculated using nonparametric tests, as appropriate. CI, confidence interval; ED, emergency department; PC, primary care. Diagnostic delays also varied according to the predominant symptom at presentation (Supplementary Table S1, available at https://doi.org/10.1016/j.esmogo.2026.100409). The longest median delays were observed in patients with weight loss (3.47 months, 95% CI 2.56-4.99), followed by those with asthenia and/or anorexia (3.28 months, 95% CI 2.92-5.16), diarrhoea (3.02 months, 95% CI 0.99-5.48), increased transaminases (2.89 months, 95% CI 1.77-3.65), pain (2.69 months, 95% CI 1.18-3.31), and dyspepsia (2.45 months, 95% CI 0.99-10.60). The shortest diagnostic delay was observed in patients presenting with jaundice (1.36 months, 95% CI 0.66-2.69). Figure 2 shows the relationship between symptom frequency and median time to diagnosis. Symptom frequency and time to diagnosis according to the predominant symptom at presentation. Bars represent the percentage of patients with each symptom (left y-axis), and the red line indicates the median time from symptom onset to diagnosis in months (right y-axis). Numeric values inside the bars correspond to symptom frequency (%), while numbers along the red line denote the median time to diagnosis. Diagnosis assessment At diagnosis, 55.8% of patients presented with metastatic disease. Imaging for diagnosis included computed tomography in 53.5% of patients, MRI in 19.8%, and echography in 18.6%, with other imaging modalities used in 2.3%. Differences were observed across countries regarding the first imaging test carried out, with ultrasound being more frequently used in Slovenia and MRI in Italy (P < 0.001). Histological confirmation was obtained by cytology and biopsy in 16.3% and 75.6% of patients, respectively, with differences observed across countries (P < 0.001); for 8.2% of patients, no data were available. Repeated sampling was required in 10.5% of patients. Overall, 77.9% of cases were discussed by a multidisciplinary tumour board (MDTB), with variation in discussion rates across countries (P < 0.001). Full details available in Supplementary Table S2, available at https://doi.org/10.1016/j.esmogo.2026.100409. Treatment and outcomes Regarding first-line treatment, 12.8% of patients did not receive any active anticancer therapy, while 43% underwent curative-intent surgery and 44.2% received chemotherapy as a first-line systemic option (Figure 3). Among patients who underwent curative-intent surgery, 21 (56.8%) received adjuvant chemotherapy [lower proportion observed in Italy (P = 0.009)]; recurrence occurred in 15 (40.5%). First-line palliative chemotherapy was administered to 61.6% of patients, including both those treated upfront with palliative intent and those who received it after recurrence. Among these patients, 69.8% experienced disease progression, and 67.6% subsequently received further lines of systemic therapy. Further details on treatment, stratified by country, are provided in Supplementary Table S3 (available at https://doi.org/10.1016/j.esmogo.2026.100409). Initial treatment approach. The proportion of patients by initial treatment approach. Systemic treatment refers exclusively to first-line therapy administered as the initial management strategy for unresectable or advanced disease; patients who received chemotherapy only after recurrence following curative-intent surgery are not included in this category. At the time of analysis, 37 patients (43.02%) had died. The median follow-up was 13.47 months (range 0.29-41.49), and the estimated median OS was 22.28 months (95% CI 17.42-34.51). Supplementary Table S4 (available at https://doi.org/10.1016/j.esmogo.2026.100409) provides the estimated median OS for each subgroup. In univariate Cox regression analysis, several variables were associated with OS and included in the multivariable model. In the multivariable analysis, male sex [hazard ratio (HR) 3.69, 95% CI 1.17-11.67, P = 0.026] and jaundice at diagnosis (HR 2.48, 95% CI 1.11-5.54, P = 0.027) were associated with shorter OS. Curative-intent surgery (HR 0.19, 95% CI 0.04-0.86, P = 0.032) and receipt of systemic therapy (HR 0.15, 95% CI 0.05-0.44, P < 0.001) were associated with longer OS. Discussion Our data highlights how the diagnosis of biliary tract tumours remains a significant clinical challenge, with a high proportion of patients (55.8%) presenting with metastatic disease at diagnosis, consistent with prior literature.5,16 Despite advances in imaging and diagnostic strategies, delays in diagnosis continue to impact the feasibility of curative-intent treatment and overall prognosis.1,3,5,9 In our study, the median time from symptom onset to diagnosis was 2.56 months, indicating diagnostic delay in this cohort. As in other studies of pancreatic cancer and BTC, our analysis found that nonspecific symptoms were associated with longer diagnostic intervals.6,13 Patients with weight loss, asthenia and/or anorexia, and diarrhoea experienced the longest delays, whereas those presenting with jaundice were diagnosed significantly quicker, which may reflect the more specific and clinically evident nature of this symptom.5,6,13 A numerically longer diagnostic delay was observed in patients initially evaluated by primary care than those presenting directly to the emergency department, although this difference did not reach statistical significance. This may support the need to increase awareness among general practitioners, specially in terms of early clinical and laboratory signs suggestive of BTC. On top of these, pathways for urgent referral for diagnostic tests should also be explored. European expert consensus emphasises the need for timely referral pathways and streamlined diagnostic algorithms to reduce such delays.4,9,17 Differences in symptom presentation across countries, such as a higher prevalence of asthenia and anorexia in Italy and a higher prevalence of jaundice in Slovenia, should be interpreted with caution and may be partially explained by variations in health care access, cultural perceptions of symptoms, and local referral practices.4,9,14 In addition, the observed differences in symptom distribution across countries may be influenced by the route of presentation. In our cohort, patients from Italy were more frequently diagnosed in primary care, whereas in Slovenia, most cases were identified in the emergency setting, with Spain showing an intermediate pattern. This difference in health care pathways may have influenced the clinical presentation at diagnosis: patients evaluated by primary care more frequently presented with nonspecific or subacute symptoms such as asthenia, anorexia, or weight loss, while those presenting to the emergency department more often presented with acute manifestations, including jaundice or abdominal pain. Although tumour location undoubtedly contributes to the development of jaundice, differences in the distribution of primary tumour sites across countries (e.g. a higher proportion of iCCA in Italy) may partly explain variations in symptom presentation, such as lower rates of jaundice, as well as differences in diagnostic approaches, including the more frequent use of MRI. The differences observed in our cohort are likely multifactorial and should be interpreted cautiously. These findings highlight the need to improve coordination between primary and hospital-based care and encourage investigation of persistent or unexplained liver test abnormalities. Our observations suggest that the primary care pathway may be a relevant setting for improvement, as earlier recognition of nonspecific or subacute symptoms could potentially reduce diagnostic delays. A notable finding in our cohort was the relatively high rate of curative-intent surgery (43%), exceeding the 20%-35% typically reported in international studies.2,3,7,18 This may reflect specific characteristics of the participating centres, including the widespread use of MDTBs, which discussed nearly 78% of cases and may have contributed to treatment decision-making.9,17 Among resected patients, 56.8% received adjuvant chemotherapy, consistent with evidence from the BILCAP trial,19 which established capecitabine as the standard adjuvant regimen for BTC. More than half of the cohort (61.6%) received palliative chemotherapy, either as first-line treatment of unresectable disease or after recurrence. Among these, 69.8% progressed, and most (67.6%) received subsequent lines of systemic therapy, reflecting treatment patterns in this cohort, which may also be influenced by the increasing availability of systemic options for BTC.20 As expected, in our survival analysis, both curative surgery and systemic therapy were independently associated with better OS.2,3,17,19,20 However, these findings must be interpreted cautiously given the study’s retrospective design and the likelihood of substantial selection bias, since patients selected for active treatment are inherently more likely to have better baseline prognosis and treatment fitness. Interestingly, in our cohort, male sex was significantly associated with shorter OS. This finding should also be interpreted with caution, as sex has not been consistently described as a strong prognostic factor16,21; it is also possible that this observation reflects sampling variability given the limited sample size and the uneven sex distribution across countries, particularly the predominance of male patients in the Italian and Slovenian cohorts. We also found that diagnostic delay itself did not significantly influence OS, although the analysis may have been underpowered to detect such an association, as previously described.13,14 While a longer time to diagnosis could, in theory, allow disease progression, our data do not demonstrate a clear association. This may suggest that clinical and biological tumour characteristics at presentation play an important role; however, these findings should be interpreted with caution given the relatively small sample size and the complexity of measuring true symptom onset. Therefore, no definitive conclusions can be drawn regarding the impact of diagnostic delay on survival in this cohort, although a relationship between diagnostic delay and survival is biologically plausible. Another interesting result was that patients presenting with jaundice, despite having the shortest diagnostic delay, had significantly worse survival outcomes. This apparent paradox may be explained by tumour-related factors, as jaundice usually reflects biliary obstruction caused by tumours located in anatomically critical sites such as the distal bile duct, hilum, or ampulla, which are often associated with more advanced or technically challenging disease due to their location. Moreover, biliary obstruction can lead to hepatic dysfunction, cholangitis, and the need for stenting procedures, all of which may compromise treatment delivery and patient condition, as previously discussed in other studies addressing stent-related complications in pancreato-biliary malignancies.22 This study has several limitations, including its retrospective design and the relatively small and imbalanced sample size across centres. Given the number of comparisons carried out, there is a potential risk of type I error, and findings should therefore be interpreted as exploratory. Additionally, the median follow-up was shorter than the estimated median OS, indicating that survival data are still immature and should be interpreted with caution. Moreover, clinical variables such as the exact date of symptom onset and subjective symptoms (e.g. anorexia) rely on medical record documentation, introducing potential recall and reporting bias. Furthermore, relevant socioeconomic or geographic factors were not systematically collected. Conclusion This multicentre, real-world study provides a descriptive overview of the diagnostic pathways and symptom presentation of BTCs across European centres. BTCs frequently present at advanced stages and are often associated with nonspecific or subacute symptoms, which may contribute to longer diagnostic intervals. The observed patterns highlight potential areas for improvement in early recognition and referral, particularly in primary care, and may help inform future studies aimed at optimizing diagnostic strategies in BTC. Acknowledgements AL mentored LC as part of the ESMO Virtual Mentorship Programme 2024. This article is based upon work from the ENS-CCA and the COST Action Precision-BTC Network, CA22125, supported by COST (European Cooperation in Science and Technology). Ethical Approval The study was conducted in accordance with the Declaration of Helsinki and approved by the ethics committee of each hospital. Patient consent was waived due to the retrospective nature of the study. Data Availability Data are not publicly available due to patient privacy and ethical restrictions but are available from the corresponding author upon reasonable request. Funding None declared. Disclosure AM declares travel and educational support from LEO Pharma, ROVI, Merck Sharp & Dohme (MSD), Merck, Incyte, and Lilly, and speaker honoraria from LEO Pharma, MSD, and Lilly. LC and VG declare no conflict of interest. AL declares travel and educational support from Ipsen, Pfizer, Bayer, AAA, SirtEx, Novartis, Mylan, Delcath Advanz Pharma, and Roche; speaker honoraria from Merck, Pfizer, Ipsen, Incyte, AAA/Novartis, QED, Servier, AstraZeneca, EISAI, Roche, Advanz Pharma, Jazz Therapeutics, and MSD; advisory and consultancy honoraria from EISAI, Nutricia, Ipsen, QED, Roche, Servier, Boston Scientific, Albireo Pharma, AstraZeneca, Boehringer Ingelheim, GENFIT, TransThera Biosciences, Taiho, MSD, Jazz Therapeutics, and Viatris; principal investigator-associated institutional funding form QED, Merck, Boehringer Ingelheim, Servier, AstraZeneca, GENFIT, Panbela Therapeutics, Novocure GmbH, Camurus AB, Albireo Pharma, Taiho, TransThera, Jazz Therapeutics, Roche, and Novartis; and is a member of the Knowledge Network and NETConnect Initiatives funded by Ipsen. Supplementary data Supplementary Data References Associated Data Supplementary Materials Supplementary Data Data Availability Statement Data are not publicly available due to patient privacy and ethical restrictions but are available from the corresponding author upon reasonable request.