Summary: survival across the disease course and in FGFR2 fusion-positive disease

Cholangiocarcinoma is usually diagnosed at an advanced stage. SEER reports 5-year relative survival for liver and intrahepatic bile duct cancer of 37.4% for localized, 13.4% for regional, and 3.6% for distant disease, and across 40,030 SEER cholangiocarcinoma cases diagnosed from 2001 to 2017 median overall survival from diagnosis was 8 months overall and 6 months for intrahepatic disease. In the first-line advanced setting, median overall survival was 12.9 months with durvalumab plus gemcitabine and cisplatin in the TOPAZ-1 three-year update and 12.7 months with pembrolizumab plus gemcitabine and cisplatin in KEYNOTE-966. After progression, median overall survival in ABC-06 was 6.2 months with FOLFOX plus active symptom control and 5.3 months with active symptom control alone.

FGFR2 fusion-positive disease has a comparatively favourable natural history. In an institutional series of 377 participants, FGFR genetic aberrations were associated with longer overall survival (37 versus 20 months, P < .001), which remained significant after excluding participants treated with FGFR inhibitors, and co-occurring TP53 and CDKN2A/B alterations were associated with shorter survival. Median cancer-specific survival was 123 versus 37 months with and without FGFR2 translocations in a FISH series, 5-year overall survival after resection of intrahepatic cholangiocarcinoma was 83% versus 32% by FGFR2 fusion status, and in a multicentre real-world cohort FGFR2 fusions or rearrangements retained a prognostic association with overall survival after exclusion of participants treated with targeted therapy. In the multicentre cohort, median overall survival was 26.5 months with and 17.0 months without FGFR2 fusions or rearrangements, with a hazard ratio of 0.54 (95% CI 0.30-0.95) at multivariate analysis, and after resection of intrahepatic cholangiocarcinoma 10-year overall survival was 46 vs. 22% by FGFR2 fusion status. Not every series found a difference: outcome did not differ significantly by FGFR2 fusion status in a Japanese RT-PCR series, and in a European routine-testing cohort median overall survival was 32.2 versus 13.6 months (HR = 0.81, p = 0.52). Participants with FGFR2 rearrangements may be less likely to benefit from first-line platinum chemotherapy, although time to progression did not differ by FGFR2 status in one sequenced cohort.

Natural history: asymptomatic early disease and late presentation

“CCAs are usually asymptomatic in early stages and, therefore, often diagnosed when the disease is already in advanced stages, which highly compromises therapeutic options, resulting in a dismal prognosis” (opens the source at this quote in a new tab)
“Despite advances in CCA awareness, knowledge, diagnosis and therapies, patient prognosis has not improved substantially in the past decade, with 5-year survival (7–20%) and tumour recurrence rates after resection still disappointing” (opens the source at this quote in a new tab)

“The prognosis is poor, with an all-stage 5-year overall survival of less than 20%.” (opens the source at this quote in a new tab)
“Patients are rarely diagnosed with early-stage disease and therefore curative surgery and adjuvant therapy is only feasible for a small proportion of patients.” (opens the source at this quote in a new tab)

Survival: SEER stage distribution and five-year relative survival, liver and intrahepatic bile duct cancer

Stage at diagnosis	Percent of cases	5-year relative survival
Localized	“45%”	“37.4%”
Regional	“23%”	“13.4%”
Distant	“21%”	“3.6%”
Unknown	“10%”	“11.6%”

“For liver and intrahepatic bile duct cancer, 44.7% are diagnosed at the local stage. The 5-year relative survival for localized liver and intrahepatic bile duct cancer is 37.4%.” (opens the source at this quote in a new tab)

Survival: median overall survival by cholangiocarcinoma subtype in SEER

“Median overall survival from diagnosis was 8, 6, 9, and 2 months for CCA, iCCA, eCCA, and CUP.” (opens the source at this quote in a new tab)
“The annual mortality rate declined from 45.2% in 2001 to 38.7% in 2016 for CCA, and from 57.1% in 2001 to 41.2% in 2016 for iCCA, but the rates for iCCA remained consistently higher than for eCCA, which generally remained stable (40.9% in 2001 vs 37.0% in 2016; Fig. 2F).” (opens the source at this quote in a new tab)

Overall survival in cholangiocarcinoma with FGFR genetic aberrations compared with FGFR wild-type disease

“Three hundred seventy-seven patients with CCA were identified, and 95 had FGFR GAs. FGFR2 GA was most common (n = 74, with 63 fusions) and seen in intrahepatic CCA. In patients with CCA, FGFR GAs occurred more frequently in younger patients (≤ 40 years; 20%) compared with older patients (> 40 years; 6.7%; P < .001), presented at an earlier stage (TNM stage I/II v III/IV: 35.8% v 22%, respectively; P = .001), and were associated with a longer OS compared with patients without FGFR GAs (37 v 20 months, respectively; P < .001). This difference remained significant after excluding 36 patients treated with FGFR inhibitors. There was no OS difference (P = .60) between CCA with FGFR2 fusions (n = 63) versus other FGFR GAs (n = 29). Patients with FGFR GAs had a better OS with FGFR-targeted therapy compared with standard treatment (P = .01). BAP1 mutation was the most common coexisting mutation without prognostic impact, whereas TP53 (P = .04) and CDKN2A/B (P = .04) were correlated with a shorter OS.” (opens the source at this quote in a new tab)
“CCA with FGFR GAs represents a unique subtype occurring in younger patients with an indolent disease course.” (opens the source at this quote in a new tab)

Cancer-specific survival with FGFR2 translocations

“The median cancer-specific survival for patients whose tumors harbored FGFR2 translocations was 123 months compared to 37 months for cases without FGFR2 translocations (P = .039).” (opens the source at this quote in a new tab)

Recurrence and death after resection by FGFR2 translocation status in a FISH series

“For the 139 cases without FGFR2 translocations, 77 patients (55%) developed metastases or local recurrence, and 99 patients (71%) died during clinical follow-up.” (opens the source at this quote in a new tab)
“Six of the 12 (50%) patients whose tumors harbored FGFR2 translocations died during clinical follow-up, and 6 patients were alive without evidence of disease. Only 3 patients (25%) developed metastases or local recurrence.” (opens the source at this quote in a new tab)
“The median cancer-specific survival interval for patients with FGFR2 translocations was estimated at 123 months (95% CI, 51-123 months) and was significantly longer than for patients without FGFR2 translocations (P = .039).” (opens the source at this quote in a new tab)
“The disease-free intervals for the 3 cases were 26 months, 63 months, and 125 months. This was also significantly longer compared to the cases without FGFR translocations (P = .007).” (opens the source at this quote in a new tab)
“In retrospective studies of patients treated with various combinations of therapy and not matched for performance status and other characteristics, it is difficult to determine the generalizability of survival data.” (opens the source at this quote in a new tab)
“Therefore, a larger study of patients who are matched for age, performance status, underlying function and other factors is warranted to appropriately determine if there truly is a prognostic difference.” (opens the source at this quote in a new tab)

Overall survival by FGFR2 fusion status in an RT-PCR series

“No significant differences in age, gender, tumor differentiation, clinical stage, and prognosis were detected between fusion-positive and -negative cases.” (opens the source at this quote in a new tab)
“The outcome was not significantly different between FGFR2 fusion-positive and -negative cases (log-rank test).” (opens the source at this quote in a new tab)

Survival after resection of intrahepatic cholangiocarcinoma by FGFR2 fusion status

“Twelve (13%) of these were found to have FGFR2 fusion, none of which were treated with FGFR targeted therapy. Patients with FGFR2 fusions were found to have a longer 5-year (83 vs. 32%, p = 0.01) and 10-year (46 vs. 22%, p = 0.04) OS. Five and 10-year DFS was also increased (68 vs. 33% p = 0.04) and (68 vs. 25 %, p = 0.02,). FGFR2 fusion status was the strongest independent factor associated with improved OS (HR 0.23, 0.09-0.62, p=0.003) and DFS (HR 0.18, 0.05-0.67, p=0.01).” (opens the source at this quote in a new tab)

Survival after resection by FGFR2 fusion status: pathological features, prognostic factors, and study limitations

“When iCCA patients positive for FGFR2 fusions were compared to patients without, FGFR2 fusion patients were found to have improved 5-year (83 vs. 32%, p = 0.010) and 10-year (46 vs. 22%, p = 0.042) OS, respectively (Fig. 2), suggesting that CCA with FGFR2 fusions may have a more indolent course.” (opens the source at this quote in a new tab)
“Overall, there were no significant differences between iCCA patients with FGFR2 fusion and wildtype FGFR in terms of vascular invasion, perineural invasion, lymph node positivity, or presence of satellite lesions.” (opens the source at this quote in a new tab)
“Perineural invasion (HR:3.23, 95% CI: 1.39–7.55, p = 0.007), increasing tumor size (HR = 1.19, 95% CI: 1.08–1.31, p < 0.001), and lymph node invasion (HR = 2.85, 95% CI: 1.37–5.92, p = 0.005) correlated significantly with increased risk of death.” (opens the source at this quote in a new tab)
“Importantly, this cohort of patients were treated prior to widespread adoption of adjuvant therapy for resected CCA, and this is reflected in the fact that adjuvant therapy for patients with FGFR2 fusion was noted in a minority of patients (adjuvant radiotherapy, n = 1, 8.3% and adjuvant systemic chemotherapy, n = 2, 16.7%). Moreover, none of the patients in this cohort received FGFR-targeted therapy.” (opens the source at this quote in a new tab)
“Our work, predating the widespread use of adjuvant therapy in resected patients suggests that the biology of disease in patients with FGFR2 fusion is more indolent, which cannot be explained by common surrogates of biology alone, such as the incidence of satellite lesions or lymph node metastases.” (opens the source at this quote in a new tab)
“Our study has several limitations including sources of bias that are inherent in retrospective studies.” (opens the source at this quote in a new tab)
“Due to small numbers of FGFR2 fusion patients, our analyses should be considered exploratory and may be subject to both Type 1 and Type 2 errors.” (opens the source at this quote in a new tab)

Prognostic association of FGFR2 fusions in a multicentre real-world cholangiocarcinoma cohort

“Among ESCAT I-III-positive cases, 42 patients had FGFR2 fusions/rearrangements and 59 had IDH1 mutations. After excluding patients treated with targeted therapies, only FGFR2 fusions/rearrangements retained a prognostic association with OS.” (opens the source at this quote in a new tab)

Overall and relapse-free survival by FGFR2 fusion status in a multicentre real-world cholangiocarcinoma cohort

“Of all the individual gene alterations, only FGFR2 fusions/rearrangements showed a significant association with longer OS (26.5 months, 95% CI 17.4-48.2 vs. 17.0 months, 95% CI 15.4-19.5; p = 0.006) at univariate assessment.” (opens the source at this quote in a new tab)
“At univariate analysis, FGFR2 alterations demonstrated a positive correlation with OS (HR 0.49, 95% CI 0.33-0.74, p = 0.006), confirmed at multivariate analysis (HR 0.54, 95% CI 0.30-0.95, p = 0.035) after adjusting for confounders, such as carbohydrate antigen 19-9 levels (>500 UI/ml) before first-line chemotherapy, liver metastases, synchronous metastatic disease and multiple sites of metastases.” (opens the source at this quote in a new tab)
“Moreover, FGFR2 fusions/rearrangements and IDH1 R132x mutations were evaluated as potential determinants of benefit from first-line platinum-based chemotherapy. No role was found for either alteration” (opens the source at this quote in a new tab)
“Seventeen patients with FGFR2-fused/rearranged and 28 patients with IDH1-mutated BTCs were included in the analysis for RFS. Median RFS was 10.3 months (95% CI 2.0-18.9) in the former” (opens the source at this quote in a new tab)
“no significant difference was observed compared to wild-type counterparts (12.3 [95% CI 11.1-14.6] months and 12.5 [95% CI 11.0-15.0] months for FGFR2 and IDH1 wild-type, respectively, both p >0.05).” (opens the source at this quote in a new tab)
“our data support a positive prognostic role for FGFR2 fusion/rearrangements, but not IDH1 mutations. However, not all published data confirm the same prognostic value of these alterations.” (opens the source at this quote in a new tab)

Survival with and without matched targeted therapy in a multicentre real-world cholangiocarcinoma cohort

“Interestingly, median OS of patients with ESCAT I-III alterations who did not receive a matched targeted treatment was similar to that of non-profiled patients (19.1 months, 95% CI 16.6-23.0 vs. 17.7 months, 95% CI 13.0-19.4, respectively).” (opens the source at this quote in a new tab)
“Of note, considering the 84 patients with available EMP showing ESCAT I-III alterations who received second-line treatment, median OS from second-line systemic therapy was significantly longer in those treated with (43.6 months, 95% CI 10.2-43.6) than in those who did not receive (10.8 months, 95% CI 7.8-14.4) a targeted agent (HR 0.48, 95% CI 0.3-0.9; p = 0.040).” (opens the source at this quote in a new tab)

Overall survival by FGFR2 fusion status in a European routine-testing cohort

“Additionally, in our cohort, patients with FGFR2 fusions exhibited a numerically longer mOS compared to WT cases (32.2 (fus) vs. 13.6 (wt) months; HR = 0.81, CI: 0.42–1.55, p = 0.52).” (opens the source at this quote in a new tab)
“This non-significant trend mirrors the significant survival benefit firmly established in other cohorts” (opens the source at this quote in a new tab)
“The lack of significance may be due to limited sample size.” (opens the source at this quote in a new tab)

Prognosis and chemotherapy response with FGFR2 rearrangements

“In smaller datasets, rearrangements and fusions involving the FGFR2 gene have been observed in 11%–17% of IHCC, and patients harboring these alterations have an improved prognosis but may be less likely to benefit from first‐line platinum chemotherapy [15, 16, 17, 18, 19, 20, 21].” (opens the source at this quote in a new tab)

“No significant difference in time to progression was noted with other commonly altered genes including IDH1, FGFR2, BAP1, ATM, ARID1A and TP53 (Figure 3, Supplementary Table 1).” (opens the source at this quote in a new tab)

Survival: first-line overall survival with durvalumab plus chemotherapy, TOPAZ-1 three-year update

“After a median 41.3 (95% CI 39.3-44.1) months' follow-up in all participants, median OS (95% CI) for durvalumab+GemCis vs. placebo+GemCis was 12.9 (11.6-14.1) vs. 11.3 (10.1-12.5) months (hazard ratio, 0.74 [95% CI 0.63-0.87]); 36-month OS rate was 14.6% vs. 6.9%, respectively.” (opens the source at this quote in a new tab)
“The 36-month Kaplan-Meier OS rate for the durvalumab arm was more than double that of the placebo arm, and the OS rate ratio for durvalumab vs. placebo increased over time (12 months: 1.15; 24 months: 1.75; 36 months: 2.12).” (opens the source at this quote in a new tab)
“In total, 88/685 (12.8%) participants were eLTS, with more eLTS in the durvalumab arm (58/341 [17.0%]) vs. the placebo arm (30/344 [8.7%]).” (opens the source at this quote in a new tab)

Survival: first-line overall survival with pembrolizumab plus chemotherapy, KEYNOTE-966

“Median overall survival was 12·7 months (95% CI 11·5-13·6) in the pembrolizumab group versus 10·9 months (9·9-11·6) in the placebo group (hazard ratio 0·83 [95% CI 0·72-0·95]; one-sided p=0·0034 [significance threshold, p=0·0200]).” (opens the source at this quote in a new tab)

Survival: second-line overall survival in ABC-06

“Median overall survival was 6·2 months (95% CI 5·4–7·6) in the ASC plus FOLFOX group versus 5·3 months (4·1–5·8) in the ASC alone group (adjusted HR 0·69 [95% CI 0·50–0·97], p=0·031; figure 2).” (opens the source at this quote in a new tab)
“The overall survival rate in the ASC alone group was 35·5% (95% CI 25·2–46·0) at 6 months and 11·4% (5·6–19·5) at 12 months, compared with 50·6% (39·3–60·9) at 6 months and 25·9% (17·0–35·8) at 12 months in the ASC plus FOLFOX group.” (opens the source at this quote in a new tab)
“Median progression-free survival was 4·0 months (95% CI 3·2–5·0; appendix p 4).” (opens the source at this quote in a new tab)
“The progression-free survival rate was 66·7% (95% CI 55·3–75·8) at 3 months, 32·1% (22·3–42·3) at 6 months, and 8·6% (3·8–16·0) at 12 months.” (opens the source at this quote in a new tab)

Survival: prognostic value of baseline tumour markers in the second-line setting

“Baseline CA19-9 data were available for 135 of 162 patients: 92 (68.1%) high, 43 (31.9%) low. High baseline CA19-9 was associated with shorter unadjusted OS [4.4 versus 6.4 months; hazard ratio (HR) 1.97, 95% confidence interval (CI) 1.33-2.93, P < 0.001] and unadjusted PFS (3.2 versus 5.0 months; HR 1.53, 95% CI 1.05-2.23, P = 0.027) in the FOLFOX arm.” (opens the source at this quote in a new tab)
“With each additional tumour marker that was elevated at baseline, median OS was numerically shorter: unadjusted median OS was 8.9 months if all tumour markers at baseline were not elevated and 7.0 months, 4.0 months, and 3.2 months in the event of having 1, 2, or 3 elevated baseline tumour markers, respectively (Figure 1).” (opens the source at this quote in a new tab)

Survival: real-world United States outcomes across lines of therapy

“All-cause mortality was 53% with a median survival of 11.5 months.” (opens the source at this quote in a new tab)
“The median OS from the index date was 11.5 months (Figure 3A & B). In total, 70.1% (n = 289), 47.8% (n = 197), 34.9% (n = 144) and 29.6% (n = 122) of patients survived 6, 12, 18 and 24 months, respectively, from the index date.” (opens the source at this quote in a new tab)
“The median OS time from first presumed BTC diagnosis was 13.3 months.” (opens the source at this quote in a new tab)
“The median time on 2LOT was 3.6 months and was longer for patients with stage III disease (4.6 months) than for patients with stage IV and stage I/II disease (3.6 months).” (opens the source at this quote in a new tab)

Mortality: United States and global mortality burden

“Liver and intrahepatic bile duct cancer is the sixth leading cause of cancer death in the United States. The death rate was 6.6 per 100,000 men and women per year based on 2020–2024 deaths, age-adjusted.” (opens the source at this quote in a new tab)

Lirafugratinib (Lyrfigtu)

Natural history, survival, and mortality