e21101 Background: Most progression-free (PFS) and overall survival (OS) curves fit exponential decay models on nonlinear regression analysis (NLRA). Log-linear curve shape varies with tumor & therapy type: straight line vs inflection to right (2 phase decay due to 2 distinct subpopulations with differing sensitivities) vs convex (downward inflection, eg due to stopping therapy that is slowing tumor growth).1 We compared SCLC to other cancers. Methods: As previously described,1 we did NLRA of published PFS & OS curves for etoposide-platinum in SCLC. We also assessed various therapies in other cancers. We calculated PFS & OS half-lives and classified curves by log-linear shape. Results: Of 18 evaluable SCLC PFS curves, 14 (78%) were highly convex & 18 (100%) were moderately or highly convex, compared to 35/888 (4%) highly convex & 186 (21%) moderately or high convex curves for other cancers (p < 0.0001). Of 24 evaluable SCLC OS curves, 12 (50%) were highly convex & 24 (100%) were moderately or highly convex, compared to 15/363 (4%) being highly convex & 87 (24%) moderately or highly convex for other cancers (p < 0.0001). In SCLC, 6 vs 4 cycles induction therapy minimally changed PFS half-life (median across studies 5.5 vs 5.2 months, p = 0.27) or OS half-life (9.5 vs 10.4 months, p = 0.06). PFS half-life correlated with OS half-life for other cancers (n 320, Spearman r 0.81, p < 0.0001) but not for SCLC (n 18, r 0.04, p = 0.86). Post progression survival (PPS) correlated positively with PFS half-life for other cancers (n 376, r 0.68, p < 0.0001) but correlated negatively for SCLC (n 18, r -0.43, p = 0.07). SCLC PFS curves had an initial downward inflection at a median of 2.9 months (where 87% of patients are still progression-free), followed by a 2nd inflection point at 4.5 months (where 71% are still progression-free). Median PFS half-life was 14.6 months prior to the 1st inflection point vs 1.9 months after the 2nd inflection point. Our analysis methods1 suggest that after the 2nd PFS inflection point, 23% of patients would progress by the next scan if scans are done every 3 weeks, while 40% would progress by the next scan if they are done every 6 weeks. Conclusions: PFS & OS curves for SCLC are more often highly convex than for other cancers since SCLC is sensitive to chemotherapy but progresses rapidly after induction therapy is completed. Giving 6 rather than 4 induction cycles makes no OS difference. New maintenance strategies are needed. Very short PFS half-lives post induction suggest that frequent follow up scans should be done for patients who are candidates for 2nd line therapy. 1. Stewart. Cancer Med 2019; 8:6871