AI-detected tumor-infiltrating lymphocytes and response to PD-1 based treatment in advanced melanoma.

9541 Background: Biomarkers to predict response to immune checkpoint inhibition (ICI)-treated melanoma are limited. This study evaluates AI-detected tumor-infiltrating lymphocytes (TILs) on pretreatment metastatic pathology specimens as a biomarker for response and survival in ICI-treated patients. Methods: Patients treated with first-line anti-PD1 ± anti-CTLA4 for advanced melanoma were retrospectively identified from 11 Dutch melanoma centers. Pre-treatment TILs were quantified on H&E stained slides using the Hover-NeXt algorithm trained on an independent melanoma dataset with 166.718 pathologist checked manually annotated cells. The average percentage of TILs per 200 µm² tumor area was calculated. The primary outcome was response to ICI per RECIST 1.1 with overall survival (OS) and progression free survival (PFS) as secondary outcomes. Univariable and multivariable logistic and Cox regression analyses assessed associations between a 10% increase in TILs present in pre-treatment metastatic slides and clinical outcomes. Multivariable analyses were adjusted for age, sex, disease stage, BRAF mutation, LDH and performance score. Objective response rate and Kaplan Meier survival analysis were stratified by TIL tertiles. Results: Metastatic melanoma specimens were available for 1246 patients, 441 received anti-PD1 + anti-CTLA4. Median TIL percentage was 10.2% (interquartile range 5.5% – 17.2%). A 10% higher baseline TIL percentage was associated with response (adjusted OR 1.39 [95% 1.21-1.58]), PFS (adjusted HR 0.87 [95% CI 0.81 – 0.94]) and OS (adjusted HR 0.84 [95% CI 0.77 – 0.93] in univariable and multivariable analysis (Table 1). Stratified analysis showed significant associations between TILs, response, and survival in both anti-PD1 monotherapy and combination therapy. Conclusions: AI-quantified TILs in pre-treatment melanoma metastases are correlated with improved response rates and survival in ICI treated patients. This correlation is independent of known clinical predictors. Outcome ICI Lowest Tertile Middle Tertile Highest Tertile Univariable OR / HR [95% CI] Multivariable OR / HR [95% CI] Response (%) All 47.9% 57.8% 64.5% 1.35 [1.21 – 1.52] 1.39 [1.21 – 1.58] Anti-PD1 45.2% 57.1% 66.7% 1.42 [1.24 – 1.64] 1.37 [1.17 – 1.60] Anti-PD1 + Anti-CTLA4 51.4% 62.2% 58.7% 1.23 [1.00 – 1.53] 1.48 [1.13 – 1.94] PFS (months) All 5.4 9.7 15.3 0.87 [0.81 – 0.93] 0.87 [0.81 – 0.94] Anti-PD1 5.4 11.8 16.5 0.85 [0.79 – 0.92] 0.89 [0.81 – 0.97] Anti-PD1 + Anti-CTLA4 5.3 11.0 10.2 0.89 [0.78 – 1.02] 0.80 [0.69 – 0.94] OS (months) All 21.4 38.4 49.2 0.81 [0.75 – 0.88] 0.84 [0.77 – 0.93] Anti-PD1 21.1 36.5 51.4 0.79 [0.71 – 0.87] 0.86 [0.77 – 0.96] Anti-PD1 + Anti-CTLA4 21.4 77.8 46.4 0.88 [0.75 – 1.03] 0.82 [0.69 – 0.97]