RESUMO
Background: Induction FOLFOX followed by PET-directed CRT prior to surgery demonstrated positive results in the CALGB 80803 study. We investigated the safety and efficacy of adding D, an anti-PD-L1 antibody, to PET-directed CRT. Methods: Patients (pts) with locally advanced esophageal/GEJ adenocarcinoma were enrolled. Pts received 2 cycles of mFOLFOX6 prior to repeat PET/CT. PET responders (≥35% reduction in SUV (PETr)) received 5-FU/capecitabine and oxaliplatin with RT to 50.4Gy, while induction PET non-responders (PETnr) received carboplatin/paclitaxel with RT. All Pts received D 1,500 mg q4W ×2 starting 2 weeks prior to CRT. Esophagectomy was planned 6-8 weeks after CRT. Pts with R0 resections received adjuvant D 1,500mg q4W ×6. The primary endpoint was the pathologic complete response (pCR) rate. Results: 36 pts were enrolled. Clinical ≥T3 disease was seen in 32 pts (88.9%, cT4 = 3) and ≥N1 in 23 (63.9%) pts. PD-L1 CPS was ≥1 in 25 (71.4%) of 35 tested with 14 (40%) ≥5. Microsatellite instability (MSI) was identified in 3 (8.3%) pts. 25 (70%) pts were PETr. Preop treatment was well tolerated with no new safety signals. Three pts had disease progression prior to surgery. pCR was identified in 8 (22.2%) pts and 22 (64.7%) had major pathologic response (MPR;ypTanyN0 + ≥90% response). Those with MSI tumors had ≥90% treatment response (1 pCR, 1: ypT1aN0 99% response, 1: ypT2N0, 90% response). 17 (73.9%) of 23 cN+ pts had ypN0 disease. MPR was associated with PD-L1 ≥1 (p = 0.03) and with a higher tumor mutational burden (TMB;p = 0.016) on MSK-IMPACT testing. Adjuvant D was commenced in 27 pts, with a median number of 6 cycles. Early discontinuation was due to risks of visits due to COVID19 (4, 15%), progressive disease (3, 11%), late surgical complications (2, 7%) and immune toxicity (1, 4%). With a median follow-up of 30 months, OS rates were 92% [95%CI: 83%-100%] and 85 % [95%CI: 74%-98%] at 12 and 24 months post induction. 12 and 24-month PFS rates were 81% [95%CI: 69%-95%] and 71% [95%CI: 58%-88%] respectively. In the 33 operated pts, 12 and 24-month disease free survival was 82% [95%CI: 70%-96%] and 78% [95%CI: 65%-94%], respectively. In addition to SUV on PET, total lesion glycolysis (TLG) was correlated with pathologic response. In cases with borderline change in SUV, TLG could predict response to treatment. One PETnr with 30.8% reduction in SUV had 88.1% reduction in TLG and pCR. Conversely, a PETr (-36.3%) who had an increase in TLG (39.3%) had only 40% treatment response on pathology. Conclusions: The addition of D to induction FOLFOX and PETdirected CRT prior to surgery is safe and appears effective with a high rate of pathologic response, as well as encouraging survival data. PD-L1 CPS≥1 and higher TMB may be associated with MPR. TLG is a novel PET variable that should be studied prospectively. Additional correlatives and comparison to a cohort treated with standard PET-directed CRT will be presented.