Definitions of Central Tumors in Radiologically Node-Negative, Early-Stage Lung Cancer for Preoperative Mediastinal Lymph Node Staging: A Dual-Institution, Multireader Study

Clinical Utility of a CT-based AI Prognostic Model for Segmentectomy in Non-Small Cell Lung Cancer

Background Currently, no tool exists for risk stratification in patients undergoing segmentectomy for non-small cell lung cancer (NSCLC). Purpose To develop and validate a deep learning (DL) prognostic model using preoperative CT scans and clinical and radiologic information for risk stratification in patients with clinical stage IA NSCLC undergoing segmentectomy. Materials and Methods In this single-center retrospective study, transfer learning of a pretrained model was performed for survival prediction in patients with clinical stage IA NSCLC who underwent lobectomy from January 2008 to March 2017. The internal set was divided into training, validation, and testing sets based on the assignments from the pretraining set. The model was tested on an independent test set of patients with clinical stage IA NSCLC who underwent segmentectomy from January 2010 to December 2017. Its prognostic performance was analyzed using the time-dependent area under the receiver operating characteristic curve (AUC), sensitivity, and specificity for freedom from recurrence (FFR) at 2 and 4 years and lung cancer-specific survival and overall survival at 4 and 6 years. The model sensitivity and specificity were compared with those of the Japan Clinical Oncology Group (JCOG) eligibility criteria for sublobar resection. Results The pretraining set included 1756 patients. Transfer learning was performed in an internal set of 730 patients (median age, 63 years [IQR, 56-70 years]; 366 male), and the segmentectomy test set included 222 patients (median age, 65 years [IQR, 58-71 years]; 114 male). The model performance for 2-year FFR was as follows: AUC, 0.86 (95% CI: 0.76, 0.96); sensitivity, 87.4% (7.17 of 8.21 patients; 95% CI: 59.4, 100); and specificity, 66.7% (136 of 204 patients; 95% CI: 60.2, 72.8). The model showed higher sensitivity for FFR than the JCOG criteria (87.4% vs 37.6% [3.08 of 8.21 patients], P = .02), with similar specificity. Conclusion The CT-based DL model identified patients at high risk among those with clinical stage IA NSCLC who underwent segmentectomy, outperforming the JCOG criteria. © RSNA, 2024 Supplemental material is available for this article.

Prognostic value of CT-defined ground-glass opacity in early-stage lung adenocarcinomas: a single-center study and meta-analysis

OBJECTIVES

The prognostic value of ground-glass opacity at preoperative chest CT scans in early-stage lung adenocarcinomas is a matter of debate. We aimed to clarify the existing evidence through a single-center, retrospective cohort study and to quantitatively summarize the body of literature by conducting a meta-analysis.

METHODS

In a retrospective cohort study, patients with clinical stage I lung adenocarcinoma were identified, and the prognostic value of ground-glass opacity was analyzed using multivariable Cox regression. Commercial artificial intelligence software was adopted as the second reader for the presence of ground-glass opacity. The primary end points were freedom from recurrence (FFR) and lung cancer-specific survival (LCSS). In a meta-analysis, we systematically searched Embase and OVID-MEDLINE up to December 30, 2021, for the studies based on the eighth-edition staging system. The pooled hazard ratios (HRs) of solid nodules (i.e., absence of ground-glass opacity) for various end points were calculated with a multi-level random effects model.

RESULTS

In a cohort of 612 patients, solid nodules were associated with worse outcomes for FFR (adjusted HR, 1.98; 95% CI: 1.17-3.51; p = 0.01) and LCSS (adjusted HR, 1.937; 95% CI: 1.002-4.065; p = 0.049). The artificial intelligence assessment and multiple sensitivity analyses revealed consistent results. The meta-analysis included 13 studies with 12,080 patients. The pooled HR of solid nodules was 2.13 (95% CI: 1.69-2.67; I2 = 30.4%) for overall survival, 2.45 (95% CI: 1.52-3.95; I2 = 0.0%) for FFR, and 2.50 (95% CI: 1.28-4.91; I2 = 30.6%) for recurrence-free survival.

CONCLUSIONS

The absence of ground-glass opacity in early-stage lung adenocarcinomas is associated with worse postoperative survival.

CLINICAL RELEVANCE STATEMENT

Early-stage lung adenocarcinomas manifesting as solid nodules at preoperative chest CT, which indicates the absence of ground-glass opacity, were associated with poor postoperative survival. There is room for improvement of the clinical T categorization in the next edition staging system.

KEY POINTS

• In a retrospective study of 612 patients with stage I lung adenocarcinoma, solid nodules were associated with shorter freedom from recurrence (adjusted hazard ratio [HR], 1.98; p = 0.01) and lung cancer-specific survival (adjusted HR, 1.937; p = 0.049). • Artificial intelligence-assessed solid nodules also showed worse prognosis (adjusted HR for freedom from recurrence, 1.94 [p = 0.01]; adjusted HR for lung cancer-specific survival, 1.93 [p = 0.04]). • In meta-analyses, the solid nodules were associated with shorter freedom from recurrence (HR, 2.45) and shorter overall survival (HR, 2.13).

Segmentectomy versus lobectomy for deep clinical T1a-bN0M0 non-small cell lung cancer

OBJECTIVES

Whether segmentectomy could yield similar oncological outcomes with lobectomy for cT1a-bN0M0 non-small cell lung cancer (NSCLC) in the deep lung parenchyma remained unclear. This study aimed to compare the long-term prognosis of segmentectomy and lobectomy for deep NSCLC.

MATERIALS AND METHODS

We retrospectively screened cT1a-bN0M0 NSCLC patients who underwent segmentectomy or lobectomy between 2012 and 2019. The 3D multiplanar reconstruction software was used to determine the tumor location. Log-rank test, Cox hazard proportional regression, and propensity score matching analyses were adopted for prognostic evaluation.

RESULTS

In total, 321 patients with segmentectomy and 239 subjects undergoing lobectomy with a median follow-up time of 48.2 months remained. All the patients had a R0 resection, and no 30-day or 90-day mortality was observed. The 5-year overall survival and disease-free survival were 99.0% and 96.6% for patients undergoing segmentectomy. No significant survival differences existed between segmentectomy and lobectomy after adjusting other factors (disease-free survival, DFS: HR = 1.20, 95%CI: 0.49-2.99, P = 0.688; overall survival, OS: HR = 1.09, 95%CI: 0.30-3.95, P = 0.892). After propensity score matching, patients with segmentectomy (n = 128) had a similar OS (P = 0.870) and DFS (P = 0.900) with those receiving lobectomy (n = 128). To further evaluate the outcomes of segmentectomy for deep lung cancer, 557 peripheral lung cancer patients who underwent segmentectomy at the same period were taken as the reference. As expected, segmentectomy for deep lesions achieved equivalent OS (P = 0.610) and DFS (P = 0.580) with the peripheral lesions.

CONCLUSIONS

Segmentectomy could achieve comparable long-term outcomes with lobectomy for deep cT1a-bN0M0 NSCLC with careful preoperative design and 3D navigation.