CT-derived visceral fat phenotypes to reveal a metabolically defined TOFI subgroup in a lung cancer screening cohort: An imaging-based framework for evaluating the BMI paradox.

Abstract

8065 Background: Body mass index (BMI) imperfectly reflects metabolic health and cannot distinguish subcutaneous from visceral fat, contributing to the long-standing “BMI paradox” in cancer, whereby higher BMI is sometimes associated with better survival. Visceral fat is metabolically active and promotes cancer progression. Individuals with normal BMI but high visceral fat, termed “Thin Outside, Fat Inside” (TOFI) are understudied in cancer populations. These TOFI patients may represent an unrecognized high-risk metabolic phenotype in cancer. We investigated whether CT-derived lower-thoracic visceral fat identifies TOFI as a high-risk subgroup for cancer mortality. Methods: Baseline CT scans from participants in the National Lung Screening Trial (NLST; n = 19,140), a lung cancer screening cohort designed to detect NSCLC, were segmented to extract subcutaneous fat, visceral fat, and vertebral levels. Because T9–T12 best captures abdominal visceral adiposity, the percentage volume of visceral fat relative to subcutaneous fat from this region was used for phenotyping. Low (<20%) and high (>40%) visceral fat groups were defined, yielding 9,887 participants. Visceral fat was combined with BMI (<25 vs ≥25) to define four phenotypes: P1 (LowBMI_LowVIS), P2 (HighBMI_LowVIS), P3 (HighBMI_HighVIS), and P_TOFI (LowBMI_HighVIS). Cancer-specific mortality was the primary endpoint, and Cox proportional hazards models were used to compare these phenotypes. Results: Distinct phenotypes demonstrated markedly different mortality risks (Table). The TOFI group exhibited the worst survival despite normal BMI, with nearly double the cancer mortality risk compared with the metabolically favorable reference group, P1 (HR 1.95; 95% CI, 1.31-2.91; p = 8.4e-04). In contrast, high-BMI groups showed only modest or no excess risk, supporting the limitation of BMI alone. Conclusions: CT-derived visceral fat phenotyping reveals a clinically relevant high-risk group TOFI, individuals who have normal BMI but significantly elevated cancer and all-cause mortality risk compared to even high-BMI patients. This finding directly exposes the BMI paradox, showing that normal BMI does not confer protection when visceral adiposity is high. Because this TOFI phenotype is undetectable by BMI, incorporating visceral fat assessment into lung cancer screening and survivorship models could enable earlier identification of high-risk TOFI patients and guide targeted prevention. Phenotype Cancer Mortality HR (95% CI) p-value All-Cause Mortality HR (95% CI) p-value P_TOFI (LowBMI_HighVIS) 1.95 (1.31-2.91) 8.4e-04 2.42 (1.84-3.18) 7.5e-11 P3 (HighBMI_HighVIS) 1.25 (0.91-1.71) 1.7e-01 1.59 (1.27-1.98) 3.7e-05 P2 (HighBMI_LowVIS) 0.92 (0.64-1.30) 6.2e-01 1.08 (0.84-1.39) 5.4e-01 P1 (LowBMI_LowVIS) Reference - - -