# Adding a Third Chemotherapy Drug Did Not Improve Survival in Advanced Biliary Tract Cancer: Results From the SWOG S1815 Phase III Trial In a large randomized phase III clinical trial called SWOG S1815, researchers tested whether adding a third chemotherapy drug (nab-paclitaxel) to the standard two-drug regimen of gemcitabine and cisplatin would help patients newly diagnosed with advanced biliary tract cancer. The triplet regimen, called GAP, did not improve overall survival: median survival was 14.0 months with GAP versus 13.6 months with the standard doublet GC (hazard ratio 0.91; P = .41). GAP did shrink tumors more often (31% response rate versus 21%, P = .03) but caused significantly more serious side effects, including a 60% versus 45% rate of severe blood-related toxicities (P = .003). The researchers concluded that triplet chemotherapy with GAP should not be routinely offered to patients with advanced biliary tract cancer. # Adding a Third Chemotherapy Drug Did Not Improve Survival in Advanced Biliary Tract Cancer: Results From the SWOG S1815 Phase III Trial ## Table of Contents - Key Points - Background: Why This Research Matters - What Are Biliary Tract Cancers? - Study Methods: How the Research Was Conducted - Who Took Part in the Study - Key Findings: Survival Results - Subgroup Findings: Gallbladder Cancer and Locally Advanced Disease - Tumor Response to Treatment - CA 19-9 Blood Marker Results - Side Effects and Safety - Clinical Implications: What This Means for Patients - Limitations: What the Study Could Not Prove - Recommendations for Patients - Frequently Asked Questions - Source Information ## Key Points - In a large randomized trial, adding nab-paclitaxel to gemcitabine and cisplatin did not improve overall survival in advanced biliary tract cancer. - Median overall survival was 14.0 months with the triplet versus 13.6 months with the standard doublet, a difference consistent with chance. - The triplet shrank tumors more often (31% vs 21% response rate) but caused significantly more severe blood-related side effects (60% vs 45%). - Seven treatment-related deaths occurred on the triplet versus one on the standard doublet; dose modifications were needed in 88% versus 78%. - The triplet should not be routinely offered; gemcitabine plus cisplatin remains a standard first-line option, and clinical trials are encouraged. ## Background: Why This Research Matters Biliary tract cancers are hard to treat. Most patients are diagnosed after the cancer has already spread or grown too large for surgery that could cure it. For more than a decade, the standard first treatment has been a two-drug chemotherapy combination called gemcitabine and cisplatin (GC). That standard treatment helps, but only modestly. On average, adding gemcitabine and cisplatin extends median overall survival (the time when half the patients are still alive) by about 1.5 months compared with older treatments. Researchers have therefore searched for ways to make the first-line treatment stronger. One promising idea was to add a third drug. A small, single-arm phase II study tested a triplet combination of gemcitabine, nab-paclitaxel, and cisplatin — shortened to **GAP** — in patients with advanced biliary tract cancer. The results looked encouraging: - Median progression-free survival (time before the cancer grew, or PFS) of 11.8 months - Overall response rate (ORR, the percentage of patients whose tumors shrank) of 45% - Disease control rate (DCR, patients whose tumors shrank or stayed stable) of 84% - Median overall survival of 19.2 months Those numbers were better than what researchers typically saw with GC alone. So the SWOG Cancer Research Network designed a large randomized phase III trial to compare GAP directly against GC. The scientific rationale was also biological. Preclinical laboratory data suggested that stroma-remodeling nanoparticles (tiny particles that can alter the dense tissue surrounding tumors) help deliver chemotherapy into "desmoplastic" tumors — cancers like pancreatic cancer and cholangiocarcinoma that build up thick, fibrous scar tissue around themselves. Nab-paclitaxel is one such nanoparticle-formulated drug, so researchers hoped it would help the other chemotherapy drugs reach the tumor more effectively. ## What Are Biliary Tract Cancers? Biliary tract cancers (BTCs) are a group of related but distinct cancers that arise in the bile ducts and gallbladder. They are **heterogeneous**, meaning they are not all the same disease — they differ in where they start, how they behave, and how they respond to treatment. The three main types are: - **Intrahepatic cholangiocarcinoma (ICC)** — cancer starting in the bile ducts inside the liver - **Extrahepatic cholangiocarcinoma (ECC)** — cancer starting in the bile ducts outside the liver - **Gallbladder carcinoma (GBC)** — cancer starting in the gallbladder Treatment options for BTC have expanded in recent years, largely because of new molecularly targeted therapies (drugs aimed at specific gene changes in the cancer) and immunotherapies (treatments that help the immune system attack cancer). Two randomized phase III trials showed that adding immune checkpoint inhibitors (ICIs — drugs that release the brakes on the immune system) to GC improved median overall survival: - From 11.5 to 12.8 months (hazard ratio 0.80) - From 10.9 to 12.7 months (hazard ratio 0.83) A **hazard ratio** compares the risk of an event (here, death) between two groups. A value below 1.0 favors the experimental treatment. Both of these results favored adding an ICI, but the benefit was modest, with median survival still around 12 months. Precision medicines are not suitable for most BTCs. That makes improving first-line treatment a top priority for this patient group. ## Study Methods: How the Research Was Conducted SWOG S1815 was a randomized, open-label, phase III trial. "Randomized" means participants were assigned to treatment groups by chance, and "open-label" means both patients and doctors knew which treatment was given. The trial was run by the SWOG Cancer Research Network and funded by the National Cancer Institute (NCI) National Clinical Trials Network, with 151 participating institutions. Participants were randomly assigned in a 2:1 ratio to GAP or GC. In other words, for every one patient assigned to the standard doublet, two were assigned to the triplet. This design gives researchers more safety and efficacy data on the experimental treatment. A dynamic balancing algorithm assigned treatments. This is a computerized method that keeps the two groups similar by balancing three factors: 1. Disease site (ICC versus ECC versus GBC) 1. Disease stage (locally advanced versus metastatic) 1. Zubrod performance status (0 versus 1) — a scale measuring how well a patient can carry out daily activities, where 0 means fully active and 1 means restricted in strenuous activity but able to do light work The trial was approved by the NCI Central Institutional Review Board, and all participants provided informed consent. The study followed the Declaration of Helsinki and applicable regulations governing clinical research. **Treatment details.** The two regimens were: - **GAP:** gemcitabine 800 mg/m², cisplatin 25 mg/m², and nab-paclitaxel 100 mg/m² given intravenously (into a vein) once per day on days 1 and 8 of a 21-day cycle - **GC:** gemcitabine 1,000 mg/m² and cisplatin 25 mg/m² given intravenously once per day on days 1 and 8 of a 21-day cycle Participants continued treatment until the cancer progressed or side effects became unacceptable. Researchers followed them until death or for 3 years after randomization, whichever came first. Whether to give white blood cell growth factor support (medication that boosts infection-fighting white cells) was left to the treating physician, and if used, it had to follow ASCO and National Comprehensive Cancer Network guidelines. **Who could join.** Patients had to have histologically or cytologically confirmed (proven by tissue or cell analysis) ICC, ECC, or GBC that was either metastatic (spread to other parts of the body) or locally advanced and unresectable (grown too far to be removed by surgery). Patients with ampullary cancer were not eligible. Additional requirements included: - No previous systemic therapy (body-wide treatment such as chemotherapy) for metastatic or locally advanced biliary cancer - No adjuvant therapy (treatment given after surgery) within 6 months of enrolling - Zubrod performance status of 0 or 1 - Adequate blood, liver, and kidney function - No history of grade 2 or higher peripheral neuropathy (nerve damage causing numbness, tingling, or pain in the hands and feet) - No active infection requiring systemic therapy **Endpoints and assessments.** The primary endpoint was overall survival (OS), defined as the time from random assignment to death from any cause. Secondary outcomes included progression-free survival (PFS), overall response rate (ORR), disease control rate (DCR), toxicity, and change in the cancer antigen CA 19-9 (a blood marker that can reflect tumor activity) from baseline to after three cycles of treatment. Tumors were evaluated every 9 weeks using RECIST version 1.1, a standardized set of rules for measuring tumor size on scans. Assessments were made by the local investigator without central review. Archival tumor tissue was collected when available for future exploratory studies. **Statistical plan.** Researchers assumed a median OS of 11.7 months for the GC arm. They designed the study to detect a hazard ratio of 0.7 — meaning a median OS of 16.7 months for GAP. This required 384 eligible participants, assuming 24 months of follow-up, 85% statistical power (an 85% chance of detecting a real difference if one exists), and a one-sided significance level of .025. The total planned sample size was 441 participants to allow for up to 13% ineligibility. The primary OS analysis included all eligible participants according to a modified intention-to-treat principle. This means everyone randomly assigned and eligible was included in survival analyses, even if they never received study treatment. OS and PFS were estimated with the Kaplan-Meier method, and differences between arms were tested with a stratified log-rank test. Hazard ratios and 95% confidence intervals (CIs) came from stratified Cox regression models. ## Who Took Part in the Study The trial met its enrollment goal. Researchers randomly assigned 452 participants between December 3, 2018, and February 15, 2021 — a 26.5-month period. Ten participants turned out to be ineligible, leaving 441 eligible and analyzable participants. The reasons for ineligibility included laboratory values outside the required range, baseline assessments not completed before random assignment, an insufficient cytology report, and prior systemic therapy. One participant withdrew consent immediately after random assignment and was not included in the analyses. Among the 441 eligible and analyzable participants: - 55% were female - 67% had intrahepatic cholangiocarcinoma (ICC) - 16% had gallbladder carcinoma (GBC) - 17% had extrahepatic cholangiocarcinoma (ECC) - 73% had metastatic disease (the rest had locally advanced disease) Baseline characteristics were well balanced between the two groups, which is what randomization is designed to achieve. The median age was 63.2 years in the GAP arm and 63.9 years in the GC arm. The youngest participant was 23.2 years old in both groups. The oldest was 88.8 years in the GAP arm and 83.6 years in the GC arm. Demographic breakdown by arm: - **Sex:** GAP — 132 males (45%), 162 females (55%); GC — 66 males (45%), 81 females (55%) - **Race (GAP versus GC):** White 244 (83%) versus 119 (81%); Black 18 (6%) versus 7 (5%); Asian 14 (5%) versus 7 (5%); Pacific Islander 1 (0%) versus 0; Native American 2 (1%) versus 1 (1%); multiracial 0 versus 1 (1%); unknown 15 (5%) versus 12 (8%) - **Hispanic ethnicity:** 30 (10%) versus 15 (10%); not Hispanic 256 (87%) versus 125 (85%); unknown 8 (3%) versus 7 (5%) - **Disease site:** GBC 46 (16%) versus 24 (16%); ICC 198 (67%) versus 99 (67%); ECC 50 (17%) versus 24 (16%) - **Disease stage:** locally advanced 77 (26%) versus 41 (28%); metastatic 217 (74%) versus 106 (72%) - **Performance status:** status 0 in 147 (50%) versus 75 (51%); status 1 in 147 (50%) versus 72 (49%) Of the 441 analyzable participants, 287 in the GAP arm and 134 in the GC arm began treatment. Twenty additional participants never started protocol treatment, so they could not be evaluated for side effects or response. Another 28 were not evaluable for response because they died or withdrew before the first disease assessment, or because their assessments were inadequate. These 48 participants were counted as nonresponders. ## Key Findings: Survival Results **The triplet did not improve survival.** Overall survival was not statistically significantly different between GAP and GC. "Not statistically significant" means the difference observed could plausibly be due to chance rather than a real treatment effect. The numbers: - Median OS with GAP: **14.0 months** (95% CI, 12.4 to 16.1 months) - Median OS with GC: **13.6 months** (95% CI, 9.7 to 16.6 months) - Hazard ratio: 0.91 (95% CI, 0.72 to 1.14); P = .41 A hazard ratio of 0.91 means the GAP group had about a 9% lower risk of death at any given point compared with the GC group, but the confidence interval crosses 1.0, and the P value of .41 tells us this difference is consistent with chance. In plain terms, the two treatments produced essentially the same survival. Survival rates at fixed time points were also similar: - At 12 months: 56% alive in the GAP arm versus 53% in the GC arm - At 24 months: 25% alive in the GAP arm versus 28% in the GC arm Progression-free survival was also essentially the same between the two arms: - Median PFS with GAP: **7.5 months** (95% CI, 6.4 to 8.5 months) - Median PFS with GC: **6.3 months** (95% CI, 4.4 to 8.2 months) - Hazard ratio: 0.89 (95% CI, 0.71 to 1.12); P = .32 Again, the confidence interval crossed 1.0, and the P value of .32 indicates the difference was not statistically significant. ## Subgroup Findings: Gallbladder Cancer and Locally Advanced Disease Researchers also looked at whether certain groups of patients did better on one treatment or the other. These "exploratory subset analyses" are considered hypothesis-generating rather than definitive, because the study was not designed or powered to answer these questions. **Gallbladder cancer.** GAP improved progression-free survival more in participants with gallbladder carcinoma than in those with intrahepatic or extrahepatic cholangiocarcinoma. This difference was statistically significant (interaction P = .01). However, the same was not true for overall survival (interaction P = .28), meaning there was no clear survival advantage for the gallbladder cancer subgroup. The specific PFS and OS figures by disease site were: - **Intrahepatic cholangiocarcinoma:** GAP median PFS 7.5 months (95% CI, 6.2 to 8.7) and OS 13.5 months (95% CI, 11.3 to 15.8); GC median PFS 7.2 months (95% CI, 5.0 to 9.5) and OS 13.6 months (95% CI, 9.5 to 19.6) - **Extrahepatic cholangiocarcinoma:** GAP median PFS 7.1 months (95% CI, 4.1 to 9.2) and OS 15.9 months (95% CI, 9.2 to 18.5); GC median PFS 7.8 months (95% CI, 4.0 to 16.8) and OS 16.3 months (95% CI, 5.1 to 29.4) - **Gallbladder carcinoma:** GAP median PFS 9.3 months (95% CI, 6.0 to 12.5) and OS 17.0 months (95% CI, 11.3 to 20.7); GC median PFS 4.1 months (95% CI, 2.8 to 6.2) and OS 9.3 months (95% CI, 7.0 to 22.2) **Disease stage.** Participants with locally advanced disease appeared to gain more from GAP than those with metastatic disease, for both overall survival and progression-free survival. But these differences were not statistically significant (interaction P = .14 for OS and P = .17 for PFS). By stage, the results were: - **Locally advanced:** GAP median PFS 9.3 months (95% CI, 8.3 to 11.5) and OS 19.2 months (95% CI, 16.3 to 24.3); GC median PFS 7.6 months (95% CI, 4.1 to 10.3) and OS 13.7 months (95% CI, 8.8 to 21.8) - **Metastatic:** GAP median PFS 6.5 months (95% CI, 5.9 to 7.9) and OS 13.0 months (95% CI, 10.6 to 14.3); GC median PFS 6.1 months (95% CI, 4.1 to 8.0) and OS 13.6 months (95% CI, 9.3 to 19.6) In short, the locally advanced subgroup had roughly a 5-month difference in median overall survival favoring GAP, but the study was not large enough to confirm that this was a real effect rather than chance. ## Tumor Response to Treatment **GAP shrank tumors more often than GC did.** Among the 290 participants with measurable disease in the GAP arm, researchers observed 6 complete responses (no detectable tumor) and 85 partial responses (tumor shrank substantially). Among the 146 participants with measurable disease in the GC arm, they observed 1 complete response and 30 partial responses. The overall response rate and disease control rate were both significantly higher with GAP: - Overall response rate: **31% with GAP versus 21% with GC** (P = .03) - Disease control rate: **78% with GAP versus 67% with GC** (P = .03) Both differences were statistically significant — the P value of .03 means there is about a 3% probability that a difference this large would appear by chance alone if the treatments were truly equal. So patients on GAP were more likely to see their tumors shrink. But that benefit did not translate into living longer, which is the key finding of the trial. ## CA 19-9 Blood Marker Results CA 19-9 is a protein that can be measured in the blood. Levels often rise when biliary tract cancer is active, so doctors sometimes use it to help track how a patient is responding to treatment. Baseline and post-cycle-3 (after three cycles of treatment) CA 19-9 levels were available for 339 of 441 eligible and analyzable participants — 77% of the group. One participant had a missing baseline CA 19-9 level, and 96 had a missing post-treatment level. Changes in CA 19-9 did not differ significantly between the treatment arms: - Median change with GAP: −12 U/mL (interquartile range, 292.31 U/mL) - Median change with GC: −3 U/mL (interquartile range, 138.38 U/mL) - P = .63 An increase in CA 19-9 was not significantly associated with response rate in the overall combined cohort (GAP n = 237 plus GC n = 102). However, that relationship did vary somewhat by treatment arm. A rise in CA 19-9 was linked to lower odds of response in the GAP arm, but no significant association appeared in the GC arm (interaction P = .09). ## Side Effects and Safety **GAP caused more serious side effects than GC.** This was one of the clearest findings of the trial. A total of 421 of the 441 eligible participants (95%) were evaluable for side effects, which researchers refer to as adverse events (AEs). The most common severe side effects — grade 3 or 4 treatment-related events occurring in at least 10% of participants overall — were blood-related problems: anemia (low red blood cell count), neutropenia (low neutrophil count, a type of white blood cell), and thrombocytopenia (low platelet count). Grade 3 or higher treatment-related blood-related side effects were significantly more common with GAP: - **60% with GAP versus 45% with GC** (P = .003) To put that in absolute terms: about 6 in 10 patients on GAP experienced a severe blood-related side effect, compared with about 4.5 in 10 on GC. Non-blood-related (nonhematologic) grade 3-4 side effects that occurred significantly more often with GAP than GC were: - Elevated ALT (a liver enzyme) - Loss of appetite (anorexia) - Constipation - Diarrhea - Swelling of the limbs (edema) - Fatigue - Low magnesium levels (hypomagnesemia) - Nausea - Sepsis (a life-threatening whole-body response to infection) - Sensory peripheral neuropathy (nerve damage causing numbness, tingling, or pain) - Vomiting Specific side effect rates from the trial included: - **Fatigue:** grade 1-2 in 173 (60%) with GAP versus 77 (57%) with GC; grade 3-4 in 27 (9%) versus 8 (6%) - **Nausea:** grade 1-2 in 148 (52%) versus 79 (59%); grade 3-4 in 11 (4%) versus 1 (1%) - **Anemia:** grade 1-2 in 132 (46%) versus 49 (37%); grade 3-4 in 95 (33%) versus 30 (22%) - **Neutrophil count decreased:** grade 1-2 in 61 (21%) versus 38 (28%); grade 3-4 in 105 (37%) versus 37 (28%) - **Platelet count decreased:** grade 1-2 in 111 (39%) versus 33 (24%); grade 3-4 in 56 (20%) versus 20 (15%) - **Peripheral sensory neuropathy:** grade 1-2 in 123 (43%) versus 30 (22%); grade 3-4 in 10 (4%) versus 1 (1%) - **Diarrhea:** grade 1-2 in 99 (35%) versus 34 (26%); grade 3-4 in 13 (5%) versus 1 (1%) - **White blood cell count decreased:** grade 1-2 in 75 (26%) versus 37 (28%); grade 3-4 in 46 (16%) versus 10 (8%) - **Lymphocyte count decreased:** grade 1-2 in 55 (19%) versus 21 (16%); grade 3-4 in 30 (11%) versus 4 (3%) Overall, the highest grade of any side effect was grade 3 or higher in 206 participants (71%) on GAP versus 81 (60%) on GC. For blood-related toxicities specifically, grade 3 or higher occurred in 173 (60%) versus 60 (45%). For non-blood-related toxicities, grade 3 or higher occurred in 118 (41%) versus 38 (28%). **Deaths related to treatment.** Seven deaths (grade 5 events, meaning fatal) occurred and were attributed to GAP: 1. Cardiac arrest (one) 1. Sepsis (three) 1. Superior vena cava syndrome — a blockage of the large vein that carries blood from the upper body to the heart (one) 1. Thromboembolic event — a blood clot (one) 1. Upper gastrointestinal hemorrhage — severe bleeding in the upper digestive tract (one) On the GC arm, only one participant experienced a fatal (grade 5) event. That death was due to progressive disease and possibly contributed to by cisplatin. **Dose reductions and stopping treatment.** Participants on GAP needed more dose modifications: - Dose modifications: **88% with GAP versus 78% with GC** (P = .008) - Stopping protocol treatment because of toxicity: 72 of 294 (24%) with GAP versus 28 of 147 (19%) with GC (P = .20) — a difference that was not statistically significant Reasons participants went off treatment were tracked in both arms. In the GAP arm: death (9), progression (134), participant refusal (30), toxicities (72), physician decision (24), and other (18). In the GC arm: death (6), progression (59), participant refusal (23), toxicities (28), physician decision (8), and other (10). ## Clinical Implications: What This Means for Patients **Triplet chemotherapy with GAP should not routinely be offered to patients with advanced biliary tract cancer.** That is the direct conclusion from the trial authors and from the Journal of Clinical Oncology associate editor who wrote the study's relevance commentary. The research team noted several important context points. This was the first randomized phase III trial conducted entirely in the United States for newly diagnosed advanced biliary tract cancers. It enrolled over 450 patients in just over 2 years, which demonstrates the large unmet need for better therapies in this population. The observed median overall survival of approximately 14 months in both arms is among the highest OS reported to date for a randomized phase III trial in biliary tract cancer. The trial also enrolled similar numbers of participants with intrahepatic cholangiocarcinoma, extrahepatic cholangiocarcinoma, and gallbladder carcinoma compared with other randomized studies, and a similar proportion of patients with locally advanced disease. So the results are broadly representative. Why might a triplet that looks promising in a phase II study fail in phase III? The authors point to history. In a previous phase II study, GAP produced a 45% response rate, with 20% of patients converting from unresectable to resectable disease (their tumors shrank enough that surgery became possible). The current trial did not replicate that high response rate, though GAP's 31% response rate was still significantly better than GC's 21%. Triplet chemotherapy combinations do have a role in treating several gastrointestinal cancers, where some regimens have shown superior overall survival versus doublet chemotherapy in randomized studies. But that has not held true for biliary tract cancer. The authors point to the phase III PRODIGE 38 AMEBICA trial, which tested a modified four-drug regimen called FOLFIRINOX (fluorouracil, leucovorin, irinotecan, and oxaliplatin) against GC in advanced biliary tract cancer. In a previous phase II study of oxaliplatin, irinotecan, and S-1 triplet chemotherapy, the response rate was 50%. In the phase III AMEBICA trial, the response rate was only 25%, with no overall survival advantage over GC alone. In fact, median overall survival was *lower* for modified FOLFIRINOX at 11.7 months (95% CI, 9.5 to 14.2) versus 13.8 months (95% CI, 10.9 to 16.1) for GC. The authors concluded that "the survival benefit of intensifying cytotoxic chemotherapy to triplet regimens in an unselected population of newly diagnosed advanced BTC remains elusive." In other words, giving more chemotherapy drugs at once has not translated into longer life for this patient group overall. For comparison, GAP's 31% response rate is similar to what immune checkpoint inhibitor combinations with GC have produced in other major trials: 26.7% in the TOPAZ-1 trial and 29% in the KEYNOTE-966 trial. But despite encouraging response trends, those trials also did not show large survival gains, with median overall survival remaining around 12 months. Since the pivotal ABC-02 study in 2010, GC has been the standard first-line systemic treatment for advanced biliary tract cancer, with a median overall survival of 11.7 months. Adding immune checkpoint inhibitors pushed that median to roughly 12 to 13 months. Molecular profiling has opened the door to targeted therapies for some patients in the second-line setting and beyond, but precision treatments are not applicable for most biliary tract cancers. The take-home message for patients and families is that more chemotherapy is not automatically better. The GAP triplet produced more tumor shrinkage but no survival benefit, and it caused meaningfully more harm — including seven treatment-related deaths, compared with one on the standard arm. ## Limitations: What the Study Could Not Prove Several limitations should be kept in mind when interpreting these results. - **Subgroup findings are exploratory.** The apparent advantages of GAP in gallbladder cancer (PFS interaction P = .01) and in locally advanced disease (interaction P = .14 for OS, P = .17 for PFS) come from analyses the study was not designed or powered to answer. These results should be considered hypotheses for future research, not established facts. - **Tumor assessments were not centrally reviewed.** Response was judged by the local investigator rather than by an independent central review panel, which can introduce variability in how responses are scored. - **CA 19-9 data were incomplete.** Only 339 of 441 participants (77%) had both baseline and post-treatment CA 19-9 levels available, which limits the strength of those analyses. - **Missing response data.** Forty-eight participants were counted as nonresponders because they died, withdrew, never started treatment, or had inadequate assessments. This could affect response rate calculations. - **Patient population considerations.** All participants had a performance status of 0 or 1, meaning they were relatively fit. Results may not apply to patients who are frailer or less able to tolerate treatment. - **No central confirmation of the primary survival endpoint was needed** (survival is an objective endpoint), but toxicity was self-reported by treating physicians per standard criteria. The authors also noted that although the triplet's toxicity was apparent, the study does not tell us whether a lower dose of nab-paclitaxel or a different schedule might have produced a better balance of benefit and harm. That would require another trial. ## Recommendations for Patients Based on this trial, here is what patients and caregivers should know: 1. **Ask about standard treatment first.** Gemcitabine plus cisplatin (GC) remains a standard first-line option for advanced biliary tract cancer. Adding a third chemotherapy drug (nab-paclitaxel) did not prolong survival in this trial. 1. **Ask about immune checkpoint inhibitors.** Adding an ICI to GC improved median overall survival by roughly 1 to 2 months in two large randomized trials. Discuss whether you are a candidate. 1. **Ask about molecular testing.** Gene testing of your tumor may reveal targetable alterations. These are relevant mostly in second-line treatment and beyond, and only in a subset of patients — but they can matter a great deal for those who have them. 1. **Ask about clinical trials.** The authors emphasize that better first-line options for advanced biliary tract cancer remain a top priority. A clinical trial may offer access to treatments not yet approved. 1. **Discuss side effect monitoring.** Both regimens cause significant side effects. GAP caused severe blood-related side effects in 60% of patients versus 45% with GC, and it required dose modifications in 88% versus 78%. If you receive either regimen, ask your care team how your blood counts and nerve symptoms will be monitored. 1. **Report neuropathy early.** Peripheral sensory neuropathy (numbness, tingling, or pain in the hands and feet) occurred in 43% of patients on GAP versus 22% on GC for mild to moderate symptoms, and in 4% versus 1% for severe symptoms. Nerve damage can be permanent if not addressed promptly. 1. **Question "more is better."** This trial shows that a larger response rate on scans does not always mean living longer. When weighing treatment options, ask about overall survival benefits and side effect risks, not just how much a tumor might shrink. ## Frequently Asked Questions ### Does adding a third chemotherapy drug help people with advanced biliary tract cancer live longer? In a large randomized trial, adding nab-paclitaxel to gemcitabine and cisplatin did not improve overall survival. Median survival was 14.0 months with the triplet versus 13.6 months with the standard doublet, a difference consistent with chance. The triplet also caused more serious side effects, so it should not be routinely offered. ### What are the side effects of the GAP triplet regimen? In the trial, 60% of patients on GAP had severe blood-related side effects versus 45% on the standard doublet. Other serious side effects included fatigue, nausea, diarrhea, nerve damage, and sepsis. Seven treatment-related deaths occurred on GAP versus one on the standard arm. Dose modifications were needed in 88% versus 78%. ### What does a hazard ratio of 0.91 mean in this trial? A hazard ratio compares the risk of death between two groups. A value of 0.91 means the GAP group had about a 9% lower risk of death at any given point. However, the confidence interval crossed 1.0 and the P value was .41, so this difference could be due to chance rather than a real effect. ### Did the triplet shrink tumors more often than the standard treatment? Yes. The overall response rate was 31% with GAP versus 21% with the standard doublet, and the disease control rate was 78% versus 67%. Both differences were statistically significant. However, this greater tumor shrinkage did not translate into longer overall survival, which was the key finding. ### What is the standard first-line treatment for advanced biliary tract cancer? Gemcitabine plus cisplatin (GC) remains a standard first-line option. Adding an immune checkpoint inhibitor to GC improved median overall survival by roughly 1 to 2 months in two large randomized trials. Molecular testing may reveal targetable alterations, mostly relevant in second-line treatment and beyond for a subset of patients. ### What should I know about nerve damage (neuropathy) from these treatments? Peripheral sensory neuropathy—numbness, tingling, or pain in hands and feet—occurred in 43% of patients on GAP versus 22% on the standard doublet for mild to moderate symptoms, and 4% versus 1% for severe symptoms. Nerve damage can be permanent if not addressed promptly, so report symptoms early to your care team. ### Should I get a second opinion before starting gemcitabine and cisplatin for advanced biliary tract cancer? A second opinion can help confirm the diagnosis and review whether standard gemcitabine plus cisplatin is the right first-line choice. Adding nab-paclitaxel did not improve overall survival (14.0 versus 13.6 months) and caused more severe blood-related side effects (60% versus 45%), so triplet chemotherapy should not be routinely offered. A second opinion can also address whether you are a candidate for an immune checkpoint inhibitor added to gemcitabine and cisplatin, or for molecular testing. Diagnostic Detectives Network provides independent expert second opinions. ## Source Information **Original article title:** SWOG S1815 - Phase III Randomized Trial of Gemcitabine, Cisplatin, and Nab-Paclitaxel in Biliary Tract Cancer **Authors:** Rachna T. Shroff, MD, MS, FASCO; Gentry King, MD; Sarah Colby, MS; Aaron J. Scott, MD; Mitesh J. Borad, MD; Laura Goff, MD; Khalid Matin, MD; Amit Mahipal, MD; Aparna Kalyan, MD; Milind M. Javle, MD; Imane El Dika, MD; Benjamin Tan, MD; Puneet Cheema, MD; Anuj Patel, MD; Renuka Iyer, MD; R. Katie Kelley, MD; Jaykumar Thumar, MD; Anthony El-Khoueiry, MD; Katherine A. Guthrie, PhD; E. Gabriela Chiorean, MD; Howard Hochster, MD; and Philip A. Philip, MD **Publication:** Journal of Clinical Oncology, Volume 43, Issue 5, pages 536-544 **DOI:** https://doi.org/10.1200/JCO-24-01383 **Accepted:** November 7, 2024. **Published:** December 13, 2024. **ClinicalTrials.gov identifier:** NCT03768414 **Funding and conduct:** Conducted by the SWOG Cancer Research Network, funded by the National Cancer Institute (NCI) National Clinical Trials Network, with 151 participating institutions. *This patient-friendly article is based on peer-reviewed research published in the Journal of Clinical Oncology. It is intended for educational purposes and should not replace discussion with your own medical team about your individual diagnosis and treatment options.* --- Publisher: Diagnostic Detectives Network (https://diagnosticdetectives.com) — independent multi-expert medical second opinions, worldwide, private-pay. Author byline: Anton Titov, MD, PhD. Contact: https://diagnosticdetectives.com/pages/contact Canonical page: https://diagnosticdetectives.com/products/adding-a-third-chemotherapy-drug-did-not-improve-survival-in-advanced-biliary-tract-cancer-results-from-the-swog-s1815-phase-iii-trial