Correlation of CT changes, PET response, and histopathologic tumor regression in treated advanced <i>EGFR</i> -mutant NSCLC before disease progression.

Abstract

8655 Background: Previous studies have examined the correlation between radiologic and pathologic responses using PET, CT, and histopathologic regression in patients receiving chemotherapy or immunotherapy. However, data on the relationship among these modalities after EGFR-TKI treatment remain limited, particularly in the context of emerging induction EGFR-TKI strategies. Methods: We conducted a two-arm, phase II clinical trial (PTR-1) enrolling patients with advanced EGFR-mutant NSCLC. After 12 weeks of EGFR-TKI therapy, participants were randomized (1:1) to either continue TKI treatment or undergo primary tumor resection. Surgery aimed to achieve locoregional control with negative margins, and lymph node dissection was not mandatory. All resected specimens were evaluated for pathologic response according to IASLC recommendations. Radiologic response was assessed by RECIST 1.1, and major pathologic response (MPR) and PET metabolic response were evaluated. Results: A total of 91 patients were enrolled, with 72 randomized into two treatment arms; 2 patients declined surgery. The median interval from treatment initiation to surgery was 3.7 months, and 18% of patients underwent resection beyond 4.5 months. Among the 34 resected cases, MPR was achieved in 29.4% (10/34) and pathologic complete response (pCR) in 5.9% (2/34). Patients harboring EGFR exon 19 deletions demonstrated higher rates of pathologic response than those with L858R mutations. The correlation between CT-based tumor regression and histopathologic response was weak (r = 0.30). Pathologic response rates were 24.5% in patients with partial response (PR) and 43.8% in those with stable disease (SD) by RECIST. The correlation coefficients between PR or SD status of the primary tumor and pathologic response were 0.15 and −0.29, respectively. Complete metabolic resolution on FDG-PET corresponded to MPR in 66% of cases. A preoperative SUVmax &lt; 2.5 was associated with a mean pathology residual viable tumor of 20.7% and a 42.8% MPR rate, while patients with ≥90% regression in PET signal achieved MPR in 66.6% of cases. Conclusions: CT-based tumor regression modestly reflects histopathologic response after EGFR-TKI induction but with poor correlation, particularly in radiologic SD cases. This represents the first trial-based dataset analyzing imaging–pathology correspondence in advanced EGFR-mutant NSCLC prior to disease progression. PET metabolic response demonstrated relatively better discrimination of major pathologic responders, though its predictive accuracy remains limited. These findings underscore the need for refined imaging biomarkers or other NGS data to identify optimal candidates for local consolidation therapy in late stage and future neoadjuvant setting in early stage strategies in EGFR-mutant NSCLC. Clinical trial information: NCT05215548 .