At the end of the 2019 ESMO annual meeting, we asked the experts what findings they thought were practice-changing and how they planned to incorporate the new findings into the care of their patients. Read on to see what they said.

Gastrointestinal Cancers

Axel Grothey, MD:
BEACON: Multicenter, randomized, open-label, 3-arm phase III trial that evaluated encorafenib plus cetuximab with or without binimetinib for patients with BRAF V600E–mutant metastatic colorectal cancer (CRC) and progressive disease after 1 or 2 previous regimens.

Key findings: Both the triplet regimen with encorafenib, binimetinib, and cetuximab and the doublet regimen with encorafenib and cetuximab led to significant improvement of median OS (triplet: 9.0 vs 5.4 months with control; HR: 0.52; P < .0001; doublet: 8.4 months; HR: 0.60; P = .0003) and ORR (triplet: 26%; doublet: 20%; control: 2%) vs standard of care (irinotecan or irinotecan-based therapy plus cetuximab) in previously treated patients with BRAF V600E–mutant mCRC. 

My take-aways: The BEACON trial really addressed a patient population with unmet needs. These patients with BRAF V600E–mutant CRC in the advanced setting have a very poor prognosis and make up approximately 8% to 12% of the patient population. Typically, patients diagnosed with this type of metastatic CRC have a median survival of approximately 1 year, and patients reaching second-line or third-line treatment have rapid progression. The primary endpoint of OS with the triplet regimen was clearly met, with an HR of 0.52 for the triplet over control, which is the largest difference ever seen in a randomized comparison in advanced CRC. That the doublet of encorafenib and cetuximab also improved OS vs control, and had a similar impact on PFS to the triplet, is quite interesting. Now the question, of course, is, do patients need a triplet regimen, or is the doublet sufficient? For the OS analysis, the number of events was insufficient to evaluate any difference between the triplet vs doublet regimen; however, there seemed to be a trend toward improved outcome with the triplet. These data will need to mature, but it is important to note that the comparison of triplet vs doublet is exploratory in nature and was not part of the initial planned analysis. Overall, the treatments were well tolerated. The triplet regimen was found to have higher adverse event rates than the doublet, primarily in terms of diarrhea. However, both triplet and doublet regimens compared favorably with the irinotecan-based control arm.

Clinical implications: Overall, the BEACON trial established a new standard of care with the BRAF/MEK/EGFR inhibitor combination that is currently listed in guidelines. I have started using this regimen in my practice. I start with the triplet combination in the second or third line. If my patients begin to experience issues from the MEK inhibitor component, such as diarrhea or blurry vision, then I remove the binimetinib component and continue with the doublet regimen of encorafenib and cetuximab. In my opinion, the data are strong enough for the doublet to take the approach of starting with the triplet regimen and moving ahead with the doublet in cases of poor tolerance. In addition, it is notable that there is an overlap in this population between MSI-H tumors and BRAF V600E–mutant tumors. If I have a patient who is MSI high and has the BRAF V600E mutation, I typically start with immunotherapy using one of the checkpoint inhibitors and use the BEACON triplet regimen in the second-line or third-line setting.

ATTRACTION-3: International, open-label, randomized phase III trial evaluating second-line nivolumab vs chemotherapy in patients with unresectable esophageal squamous cell carcinoma (ESCC) who were refractory or intolerant to 1 previous fluoropyrimidine-based and platinum-based therapy. The ATTRACTION-3 study was conducted primarily in Asia, comprising 96% of the study population, and small proportions of study centers were in Europe or the United States.

Key findings: The primary endpoint of OS was significantly improved with second-line nivolumab vs paclitaxel or docetaxel in patients with unresectable ESCC (10.9 vs 8.4 months, respectively; HR: 0.77; P = .019).

My take-aways: The observed OS benefit was independent of tumor PD-L1 expression levels. When evaluating both the OS and PFS curves, there appears to be a trend in favor of chemotherapy in the short term, followed by a crossover and benefit of nivolumab in the longer term. This suggests there is a subgroup of patients who will benefit from nivolumab. However, it is not currently clear which patients will derive the most benefit from nivolumab. Similar data have been seen with pembrolizumab in gastric cancer in the first-line and-second line settings. The ORR of nivolumab vs chemotherapy was similar at 19% vs 22%, respectively. Thus, although the response rate overall was not different between groups, the durability of response was longer with nivolumab compared with chemotherapy.

These findings show further evidence that a subgroup of patients can benefit from PD-1 antibodies in esophageal/gastric cancer, including patients with squamous cell cancer. From a safety perspective, nivolumab showed a tolerable safety profile, with numerically fewer grade 3/4 serious treatment-related adverse events vs chemotherapy (8%/2% vs 15%/4%, respectively). In addition, the prespecified exploratory analyses observed significantly improved health-related quality-of-life outcomes with nivolumab.

Clinical implications: I have started to use nivolumab in the second-line or third-line setting for patients with ESCC, particularly for those who are not ideal candidates for chemotherapy, knowing that immune checkpoint inhibitors are relatively well tolerated. I look forward to additional data on immune checkpoint inhibitors in a broader patient population to evaluate whether the findings seen here in a largely Asian population can be translated into a more diverse patient group. Trials are ongoing evaluating these questions.

ClarIDHy: International, double-blind, randomized phase III trial evaluating ivosidenib vs placebo in previously treated, advanced, IDH1-mutant positive cholangiocarcinoma.

Key findings: Ivosidenib demonstrated significantly longer median PFS vs placebo for previously treated patients with IDH1 mutation–positive advanced cholangiocarcinoma (2.7 vs 1.4 months, respectively; HR: 0.37; P < .001).

My take-aways: Using a blinded, independent radiologic review, the PFS improvement was demonstrated with a striking 63% reduction in the risk of progression or death. The disease control rate was 53% with ivosidenib vs 28% with placebo. Of interest, the PR rate with ivosidenib was only 2%, and 51% of patients had stable disease. Thus, the benefit with ivosidenib is really in stabilizing disease and preventing disease progression. Based on the crossover design, there was no statistically significant difference in OS for the overall patient population. However, in a rank-preserving structural failure time analysis used to adjust for the crossover, an improvement in OS was observed (HR: 0.46; P < .001). The safety and tolerability of ivosidenib were manageable, with improved quality of life vs placebo.

Clinical implications: I now have clinical experience using ivosidenib in this setting. I have used this agent now off-label for 3 patients, who tolerated the treatment quite well. It is too early to know whether I will see clinical benefit in my patients. One patient had early progression of disease; the other 2 are still on therapy but have not yet been restaged. Treatment options for intrahepatic cholangiocarcinomas are limited, so I can see that ivosidenib or other IDH1 inhibitors really could play a role in our standard of care.

It is worth noting that intrahepatic cholangiocarcinomas—or GI malignancies in general—have the highest yield of targetable mutations and alterations. This is the patient group where I believe we really need to have upfront molecular screening for the IDH1 mutation, FGFR fusions, and genetic aberrations to best target the disease.

Gynecologic Cancers

Bradley J. Monk, MD, FACS, FACOG:
At ESMO, reports were presented for 3 key ovarian cancer trials evaluating PARP inhibitors as maintenance therapy after first-line chemotherapy:

PRIMA: Double-blind, randomized, placebo-controlled phase III trial evaluating niraparib as maintenance therapy after initial platinum-based therapy in patients with newly diagnosed advanced ovarian cancer.

Key findings: The overall patient population receiving maintenance niraparib showed significantly longer median PFS compared with patients who received placebo (13.8 vs 8.2 months, respectively; HR 0.62; P < .001). PFS with niraparib among patients with homologous recombination deficiency (HRD) was also extended (21.9 vs 10.4 months with placebo; HR: 0.43; P < .001).

VELIA/GOG-3005: Randomized, placebo-controlled phase III trial evaluating veliparib plus carboplatin/paclitaxel followed by veliparib maintenance therapy in patients with newly diagnosed ovarian cancer.

Key findings: The addition of veliparib to platinum-based chemotherapy followed by veliparib maintenance significantly prolonged median PFS in patients with newly diagnosed high-grade serous ovarian cancer, regardless of BRCA mutation or HRD status. The risk of recurrence or progression decreased by 56% in patients with BRCA mutations, by 43% in patients with HRD, and by 32% in the overall patient population.

PAOLA-1: Randomized phase III trial for patients with newly diagnosed, FIGO stage III/IV, high-grade, serous/endometrioid ovarian, fallopian tube, or primary peritoneal cancer evaluating platinum-based chemotherapy plus bevacizumab followed by maintenance therapy with olaparib plus bevacizumab.

Key findings: Median PFS was significantly prolonged with the addition of olaparib to maintenance bevacizumab after frontline platinum-based chemotherapy (ITT patient population: 22.1 vs 16.6 months; P < .0001), regardless of BRCA mutation status or initial surgical outcome. Of note, the best PFS outcomes were seen in patients with BRCA mutations and those without BRCA mutations with HRD-positive tumors.

My take-aways: Collectively, these 3 trials reported at ESMO represent a shift in ovarian cancer first-line/maintenance therapy standard of care. Prior to these results, the standard was either platinum-based chemotherapy plus bevacizumab followed by bevacizumab maintenance or platinum-based chemotherapy followed by olaparib for patients with BRCA mutations (germline or somatic). It is not yet clear whether the expected FDA approvals based on the findings of these 3 new trials will be limited to patients with only mutated BRCA, or HRD, or will have no biomarker restriction. Based on these data, HRD is becoming the dominant biomarker in newly diagnosed, advanced ovarian cancer, and it has both prognostic and predictive capability. Among individuals with newly diagnosed ovarian cancer, homologous recombination proficiency is a negative predictor for progression and prognosis. In the PRIMA trial, for example, the patients with homologous recombination proficiency who received placebo had a median PFS of 5.4 months. However, for those with HRD, the PFS was nearly twice as long, at approximately 10 months. Consequently, the evolving development of HRD assays provides an important opportunity for predicting patient outcomes.

Clinical implications: Depending on subsequent approvals after ESMO, clinicians and patients will likely have 3 additional treatment options for newly diagnosed ovarian cancer: (1) platinum-based chemotherapy followed by maintenance niraparib; (2) platinum-based chemotherapy with veliparib followed by maintenance therapy with veliparib; and (3) platinum-based chemotherapy plus bevacizumab followed by maintenance therapy with bevacizumab plus olaparib.

Patients who started bevacizumab or veliparib as part of their initial therapy would likely continue these therapies as maintenance. However, there is also an opportunity to add a PARP inhibitor during the maintenance phase after first-line platinum-based chemotherapy. According to results from the SOLO-1 trial, olaparib maintenance therapy could be used in women with BRCA mutations, or alternatively, based on data from PRIMA, niraparib maintenance therapy could be used in any patient who received first-line chemotherapy only. In addition, for patients who received bevacizumab as part of their first-line therapy, olaparib could also be added based on the PAOLA-1 trial. The beauty of adding veliparib in the frontline setting as in the VELIA trial is the potential for all patients to try a PARP inhibitor, enabling researchers to evaluate efficacy. Even though the likelihood of response may be reduced in those patients with BRCA wild-type or homologous recombination–proficient tumors, it may be worth trying for our patients. 

KEYNOTE-146: Final primary efficacy analysis of a patient cohort with previously treated advanced endometrial cancer in an international, multicohort, open-label, single-arm phase Ib/II trial evaluating lenvatinib plus pembrolizumab.

Key findings: Clinical activity of lenvatinib plus pembrolizumab was demonstrated, independent of MSI status, PD-L1 status, or histologic subtype. Investigator-assessed ORR at data cutoff were 38.9% for all patients, 37.2% for those without MSI-H/dMMR, and 63.6% for those with MSI-H/dMMR. Responses were deep and durable, with a median DoR in all patients of 21.2 months.

My take-aways: Although PD-1 and PD-L1 agents work very well as single agents in certain cancers, including NSCLC and melanoma, they do not work effectively as single agents in the setting of non–MSI-H endometrial cancer—at least not to a clinically relevant degree. One approach to improve the efficacy of a PD-1/PD-L1 agent in this setting is to add an anti-VEGF agent, creating significant increases in complete response rates.

Clinical implications: The KEYNOTE-146 trial was transformational. Based on these results, the combination regimen of pembrolizumab plus lenvatinib, an oral VEGFR kinase inhibitor, was approved in September 2019 for the treatment of patients with advanced endometrial carcinoma that is not MSI-H or dMMR who have disease progression following previous systemic therapy and are not candidates for curative surgery or radiation.

Lung Cancer

Sandip P. Patel, MD:
CheckMate-227: Open-label, multipart, randomized phase III trial that evaluated nivolumab plus low-dose ipilimumab as first-line therapy for patients with metastatic NSCLC.

Key findings: Met the coprimary endpoint of significantly improved median OS with nivolumab plus low-dose ipilimumab compared with platinum-based doublet chemotherapy in patients with PD-L1 tumor expression ≥ 1% (17.1 vs 14.9 months, respectively; HR: 0.79; P = .007).

My take-aways: There were 2 interesting aspects to this study related to the utility of biomarkers. First, tumor mutational burden (TMB) using a cutoff of 10 mutations/Mb really did not seem to identify patients who responded to the combination of nivolumab plus low-dose ipilimumab vs those who did not any better than PD-L1 testing by IHC. Broadly, there appeared to be an OS benefit independent of either PD-L1 expression level or TMB. Of note, the greatest clinical benefit with the combination of nivolumab plus low-dose ipilimumab compared with chemotherapy appeared to be in patients with ≥ 50% PD-L1 tumor expression (HR for OS: 0.70) and those with < 1% PD-L1 tumor expression (HR for OS: 0.62). This latter finding in the patient cohort with < 1% PD-L1 expression is effectively the same survival benefit seen with pembrolizumab monotherapy in patients with high PD-L1 expression, which is particularly intriguing given that we know that PD-1 inhibition has little if any activity in patients with undetectable PD-L1 expression. What the anti–CTLA-4 therapy is doing that results in survival outcomes comparable to chemotherapy plus pembrolizumab in this group of patients with < 1% PD-L1 expression is unclear at this time. Moreover, it will be interesting to follow the durability of the survival outcomes in this study. If combining low-dose ipilimumab with nivolumab results in a similar durability in metastatic NSCLC to that reported at ESMO for patients with metastatic melanoma, I think that has the potential to be transformative.

Clinical implications: Based on these results, there are a couple of patient groups that I may consider for treatment with the combination of nivolumab plus low-dose ipilimumab, if it is approved. Putting it into clinical context, for patients with newly diagnosed metastatic NSCLC in visceral crisis who need an immediate response, I think chemotherapy plus pembrolizumab remains a standard of care independent of their PD-L1 status. For those patients who are not in visceral crisis, I would consider nivolumab plus low-dose ipilimumab in those with ≥ 50% PD-L1 expression, and especially for those patients with < 1% PD-L1 expression. Among patients with 1% to 49% PD-L1 tumor expression, I think chemotherapy plus pembrolizumab remains a strong standard of care because the data for nivolumab plus low-dose ipilimumab vs chemotherapy in this subset of patients showed equivalent median OS outcomes. Regarding the potential clinical use of nivolumab plus low-dose ipilimumab, it is important to consider that the incidence of immune-related adverse events remains relatively high, although not as high as that seen in melanoma, where a higher dose of ipilimumab is used in combination with nivolumab. Healthcare providers need to be aware of this and adept at managing these toxicities, should they occur.

IMpower110: Randomized phase III trial that evaluated single-agent atezolizumab compared with platinum-based doublet chemotherapy for chemotherapy-naive patients with metastatic NSCLC and PD-L1 expression ≥ 1% by IHC using the SP142 monoclonal antibody.

Key findings: Significantly improved median OS with atezolizumab vs chemotherapy in patients with high PD-L1 expression (TC3 or IC3) and no EGFR or ALK genetic aberrations (20.2 vs 13.1 months, respectively; HR: 0.59; P = .0106).

My take-aways: First, the PD-L1 biomarker used in this study was measured by IHC using the SP142 monoclonal antibody that is used with atezolizumab, rather than the 22C3 monoclonal antibody that is used with pembrolizumab. Of importance, distinct from the 22C3 assay, which only looks at tumor cell PD-L1 expression, the SP142 IHC assay scores PD-L1 expression in tumor cells and tumor-infiltrating immune cells. At this interim analysis with a median follow-up of 15.7 months, patients with high tumor cell (≥ 50%) and/or tumor-infiltrating immune cell (≥ 10%) PD-L1 expression, called TC3 or IC3, had a 41% reduction in the risk of death with atezolizumab vs platinum-based doublet chemotherapy. 

Clinical implications: If atezolizumab monotherapy is approved in the first-line setting for chemotherapy-naive patients with advanced NSCLC, it could represent another potential treatment option for those patients with high tumor-associated PD-L1 expression as determined with the appropriate companion assay. Of note, this would include those patients with IC3 who may not have been eligible for pembrolizumab monotherapy due to insufficient tumor cell PD-L1 expression.

FLAURA: Double-blind, randomized phase III trial comparing osimertinib vs a first-generation EGFR TKI (erlotinib or gefitinib) in previously untreated patients with advanced NSCLC adenocarcinoma.

Key findings: Final analysis of OS showed a 20% reduction in risk of death with osimertinib compared with erlotinib or gefitinib as first-line therapy (median OS of 38.6 vs 31.8 months; HR 0.799; P = .0462)

My take-aways: First, a median OS of more than 3 years with osimertinib in metastatic NSCLC is impressive and a testament to drug development. The second important finding involves the idea of whether it is better to use osimertinib as initial therapy or to sequence it after another EGFR TKI. Even in the context of this well-done study with an ideal clinical management scenario, only approximately 30% of patients with progression on erlotinib or gefitinib were able to cross over to osimertinib, even with EGFR T790M mutation testing being available. Another 30% of patients did not receive any further treatment at disease progression on erlotinib or gefitinib. I think that this last observation should serve as a warning to clinicians that, particularly outside of the setting a carefully controlled clinical trial, the percentage of patients who may not receive further treatment at disease progression is high. Thus, in the “real world,” more than one third of patients may be deprived of the best available therapy if osimertinib is saved for later lines of therapy. Given that we currently have no way to identify those patients who may have rapid disease progression vs those who may have a more indolent disease course, we cannot predict which patients may most benefit from osimertinib compared with a first-generation EGFR TKI.

Clinical implications: These results confirm the role of osimertinib as the preferred first-line therapy for patients with advanced NSCLC and an EGFR mutation wherever it is available. Overall, osimertinib therapy prolongs survival, with better CNS activity and better tolerability compared with a first-generation EGFR TKI.

Remember to download the ESMO 2019 highlights slideset from the CCO Web site!

What studies most excited you at ESMO 2019 and why? Leave your comments below to join in the conversation.