What to do when a patient presents with resectable disease?
Cross-trial comparisons are fraught with danger. Until we have a phase 3 study comparing concurrent neoadjuvant chemo/immunotherapy with concurrent adjuvant chemo/immunotherapy, I do not think we can answer the question “which is better?” However, there are some caveats to keep in mind when deciding on which approach to recommend to our patients: First, neoadjuvant treatment requires biomarker testing to ensure the patient does not have EGFR or ALK mutations. This will necessitate a delay in the operation. Will patients be willing to wait? Will the surgeon? Or, would patients prefer to proceed with surgery while the results are pending? Yes, neoadjuvant therapy gives you information regarding the pCR rate, but does that help you in subsequent management of the patient? We do not know.
Secondly, the two adjuvant studies used adjuvant chemotherapy followed by adjuvant immunotherapy, as contrasted to the neoadjuvant study which used concurrent chemo/immunotherapy. Given the longer duration of treatment in postoperative sequential adjuvant studies, there tends to be more drop off because of patients being unwilling or unfit postoperatively to receive long courses of therapy. In IMpower 010, 1,269 patients completed adjuvant chemotherapy; 1,005 were randomized, and of the 507 assigned to the atezolizumab/chemo group, only 323 completed treatment.
Finally, we must beware of using neoadjuvant chemo/immunotherapy to “down-stage” a patient. KEYNOTE-091 included patients with IIIA disease and no benefit to adjuvant chemotherapy followed by immunotherapy was found in this subgroup of patients, which leads me to wonder if these patients were appropriately selected as surgical candidates. In the NADIM II trials, 9 of 29 patients on the neoadjuvant chemotherapy were not resected.
So, many questions remain. In addition to the ones we’ve raised, there is a clear and immediate need for predictive and prognostic biomarkers. In the NADIM II trial, PD-L1 expression was a predictive biomarker of response. The pCR rate for patients with a PD-L1 tumor expression of less than 1%, 1%-49%, and 50% or higher was 15%, 41.7%, and 61.1%, respectively. However, in KEYNOTE-091, the benefit was not seen in patients with PD-L1 of at least than 50%, where the 18-month DFS rate was 71.7% in the pembro arm and 70.2% in the placebo arm.
Another possible biomarker: circulating tumor DNA. In the first NADIM study, three low pretreatment levels of ctDNA were significantly associated with improved progression-free survival and overall survival (HR, 0.20 and HR, 0.07, respectively). Although clinical response did not predict survival outcomes, undetectable ctDNA levels after neoadjuvant treatment were significantly associated with progression-free survival and overall survival (HR, 0.26 and HR0.04, respectively). Similarly, in CheckMate 816, clearance of ctDNA was associated with longer EFS in patients with ctDNA clearance than in those without ctDNA clearance in both the nivolumab/chemotherapy group (HR, 0.60) and the chemotherapy-alone group (HR, 0.63).
Hopefully, ASCO 2023 will provide more answers.
Dr. Schiller is a medical oncologist and founding member of Oncologists United for Climate and Health. She is a former board member of the International Association for the Study of Lung Cancer and a current board member of the Lung Cancer Research Foundation.