Risk stratification for lung nodules: Size isn't everything

National Trends and Outcomes of Segmentectomy in the Society of Thoracic Surgeons Database

BACKGROUND

Segmentectomy is gaining popularity as a parenchyma-sparing alternative for anatomic lung resection. This study sought to investigate temporal changes in patient selection, case volume, and outcomes for segmentectomy using the Society of Thoracic Surgeons (STS) National Database.

METHODS

The STS General Thoracic Database was queried for patients who had undergone segmentectomy as the primary procedure between 2002 and 2018. The American College of Surgeons Oncology Group definition of high-risk patients on the basis of pulmonary function and major cardiovascular comorbidities was applied. Annual trends of case volume, patient risk profile, surgical indication, approach, and outcomes were analyzed.

RESULTS

A total of 10 629 patients were analyzed from 310 contributing centers. The annual segmentectomy volume more than doubled from <4 per center in 2009 to 8.6 per center by 2017. Lung cancer was the most common indication (70.1%), followed by benign disease (15.6%) and metastatic tumors (14.3%). Although the operative indication remained constant, the subset of high-risk patients (24.5%) decreased gradually over time (slope, -0.6% per year; P = .001). After 2012, segmentectomies were most commonly performed minimally invasively (video-assisted thoracoscopic surgery, 58.3%; robotic surgery, 19.4%), with a steadily declining use of thoracotomy (overall, 22.3%; slope, -2.4%/y; P = .001). Overall complication rates decreased over the study period from 41.7% to 26.1% (slope, -0.57%/y; P = .001). The overall major complication rate was 4.6% (range, 2.0%-7.1%), 30-day mortality was 1.0% (range, 0.7%-5.0%), and both have been trending downward since 2009 (P = .01).

CONCLUSIONS

Segmentectomies are increasingly performed nationally, with a steady decline in the subset of high-risk patients over time. Complication rates have decreased significantly, parallel to the increasing use of minimally invasive techniques by video-assisted thoracoscopic surgery and robotic surgery.

Consecutive Serial Non-Contrast CT Scan-Based Deep Learning Model Facilitates the Prediction of Tumor Invasiveness of Ground-Glass Nodules

Introduction

Tumors are continuously evolving biological systems which can be monitored by medical imaging. Previous studies only focus on single timepoint images, whether the performance could be further improved by using serial noncontrast CT imaging obtained during nodule follow-up management remains unclear. In this study, we evaluated DL model for predicting tumor invasiveness of GGNs through analyzing time series CT images.

Methods

A total of 168 pathologically confirmed GGN cases (48 noninvasive lesions and 120 invasive lesions) were retrospectively collected and randomly assigned to the development dataset (n = 123) and independent testing dataset (n = 45). All patients underwent consecutive noncontrast CT examinations, and the baseline CT and 3-month follow-up CT images were collected. The gross region of interest (ROI) patches containing only tumor region and the full ROI patches including both tumor and peritumor regions were cropped from CT images. A baseline model was built on the image features and demographic features. Four DL models were proposed: two single-DL model using gross ROI (model 1) or full ROI patches (model 3) from baseline CT images, and two serial-DL models using gross ROI (model 2) or full ROI patches (model 4) from consecutive CT images (baseline scan and 3-month follow-up scan). In addition, a combined model integrating serial full ROI patches and clinical information was also constructed. The performance of these predictive models was assessed with respect to discrimination and clinical usefulness.

Results

The area under the curve (AUC) of the baseline model, models 1, 2, 3, and 4 were 0.562 [(95% confidence interval (C)], 0.406~0.710), 0.693 (95% CI, 0.538-0.822), 0.787 (95% CI, 0.639-0.895), 0.727 (95% CI, 0.573-0.849), and 0.811 (95% CI, 0.667-0.912) in the independent testing dataset, respectively. The results indicated that the peritumor region had potential to contribute to tumor invasiveness prediction, and the model performance was further improved by integrating imaging scans at multiple timepoints. Furthermore, the combined model showed best discrimination ability, with AUC, sensitivity, specificity, and accuracy achieving 0.831 (95% CI, 0.690-0.926), 86.7%, 73.3%, and 82.2%, respectively.

Conclusion

The DL model integrating full ROIs from serial CT images shows improved predictive performance in differentiating noninvasive from invasive GGNs than the model using only baseline CT images, which could benefit the clinical management of GGNs.

Impact of tumor disappearance ratio on the prognosis of lung adenocarcinoma ≤2 cm in size: A retrospective cohort study

BACKGROUND/PURPOSE

Lung cancer patients can have advanced-stages at diagnosis, even the tumor size is ≤2 cm. We aimed to study the relationship between image characteristics, clinical, and patholoigcal results.

METHODS

We retrospectively enrolled patients with lung adenocarcinoma at Taichung Veterans General Hospital and Chang Gung Memorial Hospital from 2007 to 2015, who were diagnosed with treatment naïve primary tumor lesions at sizes less than 2 cm, as measured by computed tomography (CT) scans. The patient was analyzed for lymph node (LN) and distant metastasis evaluation, with clinicopathological characteristics, including tumor-disappearance ratio (TDR) (tumor diameter at the mediastinal/lung window) over chest CT scans, pathological diagnosis, disease-free survival (DFS), and overall survival (OS).

RESULTS

Totally 280 patients were surveyed initially and showed significantly increase of clinical LN involvement and distant metastasis when TDR ≤75% compared with >75% (21.6% vs 0% for LN involvement; 27.1% vs 0% for distant metastasis; both p < 0.001). We included 199 patients having surgical treatment and follow-up for the survival analysis. With a TDR ≤75%, significantly worse DFS (HR, 19.23; 95% CI, 2.60-142.01; p = 0.004) and a trend of worse OS (HR, 4.97; 95% CI, 0.61-40.61; p = 0.134) were noted by Kaplan-Meier method. TDR ≤75% revealed more advanced pathological stage, and more tumors containing micropapillary or solid subtypes when diagnosed adenocarcinoma.

CONCLUSION

For lung cancer patients with primary tumor ≤2 cm, TDR ≤75% was related to more advanced stages, the presence of micropapillary or solid components of adenocarcinoma subtypes, worse DFS, and a trend of worse OS.

Histologic subtyping in pathologic stage I-IIA lung adenocarcinoma provides risk-based stratification for surveillance

Background

We hypothesize that recurrence hazard following resection for stage I-IIA lung adenocarcinoma (ADC) varies according to histologic subtype, which may provide risk stratification for surveillance better than the current uniform follow-up protocol.

Results

Presence (≥5%) of high-grade histologic subtypes (MIP and/or SOL) was associated with a significantly higher recurrence hazard: (1) presence of either MIP or SOL was associated with a significant increase in recurrence hazard during the first two years after surgery; (2) presence of SOL was associated with an increase in recurrence hazard—in particular, distant recurrence hazard—during the first year after surgery; (3) absence of high-grade subtypes (515/1,572 patients) was associated with a very low recurrence hazard (<2% risk/year) during the first ten years after surgery.

Methods

All hematoxylin and eosin–stained tumor slides from pathologic stage I-IIA lung ADC (n = 1572) were reviewed for quantification of the percentage of each histological subtype. Recurrence hazard was estimated using the Kernel-Epanechnikov smoothing procedure. The association between recurrence hazard and high-grade histologic subtypes (micropapillary [MIP] and solid [SOL]) was assessed.

Conclusions

Our findings suggest that histologic subtyping has utility for identifying recurrence hazard for surgically resected stage I-IIA lung ADC patients and provide rationale for establishing risk-based surveillance.