Biomedical Profiling of Lung Tumor via Ventilation-Induced Tumor Deformation: Implications on the Prognosis of Lung Cancer

Abstract

<h3>Purpose/Objective(s)</h3> To quantitatively characterize tumor biomechanical rigidity via ventilation-induced tumor deformation (VITD) and assess its prognostic values. <h3>Materials/Methods</h3> This study utilized 10-bin 4D-CT and PET scans for 16 lung cancer patients (7 squamous cell carcinoma (SCC) and 9 adenocarcinoma (AC)). First, tumor surface models were generated based on binary segmentation for 10 respiratory phases and regionally aligned using surface-based registration. The resultant shape descriptors () across respiration phases were assembled as an intra-subject VITD. Subsequently, tumor strain was approximate as the magnitude of the averaged shape changes (<i>VITD_mag</i>) between end-of-exhale (EOE) and end-of-inhale (EOI) phases in the relevant regions - areas showing significant morphological correlations with the respective lung volume trajectory. Afterward, para-tumor stress (PTS) was approximated as the averaged displacements in the dilated para-tumor region, by deforming EOE to EOI 4D-CTs. Finally, <i>VITD_rigi</i> was calculated as PTS over <i>VITD_mag</i>. To assess the prognostic value of the <i>VITD</i> model, <i>VITD_mag</i> and <i>VITD_rigi</i> were correlated with PET-derived <i>SUV_max</i> for AC and SCC tumors, respectively. <h3>Results</h3> On lung SCC tumors, <i>VITD_mag</i>, the extent of maximal shape changes in respiration, was negatively associated with <i>SUV_max,</i> indictive of elevated tumor metabolism (r= -0.88, p=0.01). After controlling for PTS, an increase in the derived tumor rigidity (<i>VITD_rigi</i>) was found to be significantly associated with increased <i>SUV_max</i>, with r=0.76 and p=0.04. In contrast, on AC tumors, featuring surfactant overproduction and thus complex tissue composition, neither <i>VITD_mag</i> nor <i>VITD_rigi</i> presented a significant association with <i>SUV_max</i> (p = 0.98 and 0.88). <h3>Conclusion</h3> We introduced a novel noninvasive pipeline to characterize lung tumor rigidity based on VITD in 4D-CT. The derived <i>VITD_rigi</i> in lung SCC significantly correlated with tumor metabolism, indicating tumor aggressiveness. The proposed VITD model provides new biomechanical evidence in lung SCC prognosis. For AC tumors with complex tissue composition, deeper radiographic profiling besides tumor deformation is needed to effectively characterize its biomechanical properties.