1125P - Accuracy of PET-CT to assess extent of nodal disease in clinical stage III melanoma

Background

With effective systemic therapies for melanoma, patients (pts) with clinical stage III (cS3) disease, previously first treated with operation, are now increasingly offered neoadjuvant therapy (NST). Excellent pathologic response rates (pRR) and improved event-free survival with NST have now raised questions about the extent of operation required for those with a favorable pRR, particularly pts with limited disease at baseline. PET-CT, for baseline and post-NST staging, is used primarily to exclude distant disease. While data exist showing nodal metabolic response is unreliable for predicting a pathologic complete response (pCR), data specifically evaluating nodal imaging to nodal pathology correlation are scarce. We evaluated correlation between PET-CT nodal findings and operative pathology.

Methods

With IRB approval, we identified cS3 melanoma pts treated with NST 2011-2024 and operated on at our institution. We included pts with both pre- and post-NST PET-CT imaging. Imaging response was reported per RECIST v1.1 and EORTC PET response criteria. Pathology assessment was per IMNC guidelines.

Results

Of 60 pts, median age was 61 years, 50% were female. NST was immunotherapy (IO) + targeted therapy (IO+TT) in 48% and IO in 42%. Median NST duration was 12 weeks. Imaging response rate (iRR) overall was 60% per RECIST and 57% per EORTC PET and was higher for IO+TT (72%) than IO (48%). Pathology evaluation of a median 15 nodes (LNs)/pt showed pCR or near-pCR (≤10% viable tumor) in 28/60 (47%) with 65% having a pRR with <50% viable tumor. Among pts with residual disease, the median number of positive LNs at operation was 2 (IQR 1-4). At baseline, 42% of pts had 1 positive LN, with a 56% iRR and 48% pRR with a 44% pCR/near-pCR rate, while 58% of pts had ≥2 positive LN with a 63% iRR and 66% pRR with 46% having a pCR/near-pCR. Enumerating the number of LN+ on PET-CT underestimated extent of disease: 27% had a greater number of involved LNs on pathology assessment than on preoperative imaging (32% for pts with 1 LN+, 23% for pts with ≥2 LN+ nodes on baseline imaging).

Conclusions

PET-CT nodal imaging should be interpreted cautiously as it may underestimate the number of affected LNs. Over-reliance on imaging to guide surgical management following NST may lead to suboptimal treatment.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

The authors.

Funding

Has not received any funding.

Disclosure

R. Stoff: Financial Interests, Personal, Invited Speaker: Merck Serono, MSD, Medison Pharma, Novartis. M.S. Block: Financial Interests, Institutional, Research Funding: BMS, Alkermes, Genentech, Merck, nFerence, Pharmacyclics, Regeneron, Transgene; Financial Interests, Personal and Institutional, Research Funding, Non-paid advisory bord member: Marker Therapeutics, Sorrento, TILT biotherapeutics, Viewpoint molecular therapeutics. T.J. Hieken: Financial Interests, Institutional, Research Funding: Genentech, SkylineDX. All other authors have declared no conflicts of interest.