(P 167) FDG-PET-CT RESPONSE AFTER NEOADJUVANT CHEMOTHERAPY PREDICTS LONG-TERM OUTCOME IN PATIENTS WITH SOFT TISSUE SARCOMA -- 15-YEAR FOLLOW-UP

Objective: Only 30 to 40% of high-grade sarcomas are responsive to initial chemotherapy (CX). CX has significant toxicities, and the evaluation of response by CT imaging has well-known limitations. FDG PET-CT has the advantage of assessing tumor metabolism in addition to tumor size changes. This prospective study evaluated whether early SUV-max changes on FDG PET-CT after neoadjuvant CX could predict outcomes in patients with high-grade sarcomas.

Methods: This prospective trial was designed to determine if the change in the SUV-max on PET after 1 or 4 cycles of chemotherapy could predict progression-free survival (PFS), overall survival (OS), and histologic response in patients with high-grade soft-tissue sarcomas receiving neoadjuvant CX with doxorubicin (free or pegylated-liposomal doxorubicin) and ifosfamide. We calculated a sample size of 56 patients to achieve 90% power to test the hazard ratio using a one-sided log-rank test at alpha = 5%. PET metabolic response was defined as the SUV-max change from baseline to following either cycle 1 of CX (initial response/Delta-1) or cycle 4 of CX before surgical excision (multiple cycle response/Delta-2) on an intent-to-treat basis. A positive responder was defined as a reduction in SUV-max reduction of >40% from baseline. Results were also analyzed using SUV-max as a continuous variable.

Results: Table 1 shows demographic characteristics. Sixty-four patients had evaluable data for Delta-1, and 56 had received additional chemotherapy cycles and were thus evaluable for a Delta-2 assessment. The long-term follow-up data was updated in May 2021 with a median follow-up time of 8.1 years (Range 0.4 to 14.9)
The 5-year PFS for patients with positive Delta-1 response was 65% vs. 34% for non-responders (p 0.021). This remained true after all chemotherapy cycles had been completed. Five-year PFS for Delta-2 responders was 63% vs. 26% in non-responders (p=0.002). Among the group of non-metastatic disease patients, 5-year PFS for Delta-1 responders was 72% vs. 42% (p=0.023) in non-responders, and 5-year PFS for Delta-2 responders 65% vs. 38% in non-responders (p=0.045). After Cox regression analysis, we found that PET-CT response was independently highly predictive of PFS. This was true regardless of whether the response was considered a continuous or categorical variable.
Similarly, a decrease in SUV-max predicted OS and correlated with histological response. 5-yr OS in Delta-1 responders was 73% vs. 56% in non-responders (p 0.1), and in Delta-2 responders 88% vs. 30% in non-responders (p < 0.001). For histological response, the correlation with Delta-1 response was 0.29 (p=0.026), n=59, and for Delta-2, the correlation was 0.51 (p < 0.001), n=52.
Importantly, long-term adverse events were noted. Two patients developed acute renal failure requiring dialysis and were left with chronic kidney disease. Two patients developed myelodysplastic syndrome and died from the disease. Three patients developed secondary malignancy (1 breast cancer, 1 radiation-induced chondrosarcoma, and 1 melanoma), and 1 developed autoimmune hemolytic anemia (AIHA).

Conclusion: This study demonstrates that the change in PET-CT SUV-max significantly predicts long-term outcomes in patients with high-grade sarcomas undergoing neoadjuvant anthracycline-based CX. This data supports using PET-CT early in treatment to assess response to anthracycline-based CX in patients with high-grade soft tissue sarcomas. In practice, assessing the response to CX with early PET-CT adds information to the decision to continue chemotherapy which is often challenging.
Finally, in our long-term follow-up analysis, 5 out of 64 patients developed secondary malignancies. This suggests that secondary malignancies and other serious adverse events after perioperative CX may be higher than captured previously by studies with shorter follow-ups.