CALGB 140503: A Randomized Phase III Trial of Lobectomy versus Sublobar Resection for Small (< 2cm) Peripheral Non-Small Cell Lung Cancer

Predicting pathological highly invasive lung cancer from preoperative [18F]FDG PET/CT with multiple machine learning models

PURPOSE

The efficacy of sublobar resection of primary lung cancer have been proven in recent years. However, sublobar resection for highly invasive lung cancer increases local recurrence. We developed and validated multiple machine learning models predicting pathological invasiveness of lung cancer based on preoperative [¹⁸F]fluorodeoxyglucose (FDG) positron emission tomography (PET) and computed tomography (CT) radiomic features.

METHODS

Overall, 873 patients who underwent lobectomy or segmentectomy for primary lung cancer were enrolled. Radiomics features were extracted from preoperative PET/CT images with the PyRadiomics package. Seven machine learning models and an ensemble of all models (ENS) were evaluated after 100 iterations. In addition, the probability of highly invasive lung cancer was calculated in a nested cross-validation to assess the calibration plot and clinical usefulness and to compare to consolidation tumour ratio (CTR) on CT images, one of the generally used diagnostic criteria.

RESULTS

In the training set, when PET and CT features were combined, all models achieved an area under the curve (AUC) of ≥ 0.880. In the test set, ENS showed the highest mean AUC of 0.880 and smallest standard deviation of 0.0165, and when the cutoff was 0.5, accuracy of 0.804, F1 of 0.851, precision of 0.821, and recall of 0.885. In the nested cross-validation, the AUC of 0.882 (95% CI: 0.860-0.905) showed a high discriminative ability, and the calibration plot indicated consistency with a Brier score of 0.131. A decision curve analysis showed that the ENS was valid with a threshold probability ranging from 3 to 98%. Accuracy showed an improvement of more than 8% over the CTR.

CONCLUSION

The machine learning model based on preoperative [¹⁸F]FDG PET/CT images was able to predict pathological highly invasive lung cancer with high discriminative ability and stability. The calibration plot showed good consistency, suggesting its usefulness in quantitative risk assessment.

Adequate Number of Lymph Nodes Sampled May Determine Appropriate Surgical Modality for Early-Stage NSCLC: A Population-Based Real-World Study

BACKGROUND

The standard surgical procedure for ≤ 2 cm non-small cell lung cancer (NSCLC), including the number of lymph nodes sampled (nLN) and surgical modality, remains controversial. This study was designed to determine the optimal cohort in which sublobectomy could be an alternative to lobectomy.

MATERIALS (OR PATIENTS) AND METHODS

Patients from 1998 to 2017 were identified from the Surveillance, Epidemiology, and End Results (SEER) database. The optimal cutoff value of nLN was identified using a restrictive cubic spline graph (RCS). Kaplan-Meier analysis was used to determine cancer-specific survival (CSS). The COX proportional hazard regression model was used to identify the influence of clinical and demographic variables on survival, and propensity score matching (PSM) was used to balance differences in baseline characteristics. Finally, we used an external cohort from a single-center medical institution to verify the conclusions drawn from the SEER database.

RESULTS

A total of 6150 patients were included. The sublobectomy subgroup included segmentectomy (308, 5.0%) and wedge resection (1611, 26.2%). The cutoff value for nLN was 7. In the nLN ≥7 subgroup of the PSM cohort, the CSS of segmentectomy and wedge resection was close to that of the lobectomy subgroup (P = .12), whereas in the nLN <7 subgroup, the CSS of the lobectomy subgroup was significantly higher than that of the sublobectomy with P < .001). Surgical methods, nLN, age, sex, and differentiated grade were independent predictors of CSS. External cohort validation: A total of 1106 patients from the Affiliated Jinhua Hospital of Zhejiang University School of Medicine between 2013 and 2020 were included. The grouping criteria were consistent with the SEER database. In the nLN≥7 subgroup, sublobectomy had a survival outcome similar to that of lobectomy (P = .81).

CONCLUSION

Sublobectomy and nLN < 7 were strongly associated with poorer CSS for early-stage NSCLC. On the premise of nLN ≥ 7, sublobectomy could provide similar survival outcomes to lobectomy for these patients.

DNA methylation patterns at and beyond the histological margin of early-stage invasive lung adenocarcinoma radiologically manifested as pure ground-glass opacity

Background

Early-stage lung cancers radiologically manifested as ground-glass opacities (GGOs) have been increasingly identified, among which pure GGO (pGGO) has a good prognosis after local resection. However, the optimal surgical margin is still under debate. Precancerous lesions exist in tumor-adjacent tissues beyond the histological margin. However, potential precancerous epigenetic variation patterns beyond the histological margin of pGGO are yet to be discovered and described.

Results

A genome-wide high-resolution DNA methylation analysis was performed on samples collected from 15 pGGO at tumor core (TC), tumor edge (TE), para-tumor tissues at the 5 mm, 10 mm, 15 mm, 20 mm beyond the tumor, and peripheral normal (PN) tissue. TC and TE were tested with the same genetic alterations, which were also observed in histologically normal tissue at 5 mm in two patients with lower mutation allele frequency. According to the difference of methylation profiles between PN samples, 2284 methylation haplotype blocks (MHBs), 1657 differentially methylated CpG sites (DMCs), and 713 differentially methylated regions (DMRs) were identified using reduced representation bisulfite sequencing (RRBS). Two different patterns of methylation markers were observed: Steep (S) markers sharply changed at 5 mm beyond the histological margin, and Gradual (G) markers changed gradually from TC to PN. S markers composed 86.2% of the tumor-related methylation markers, and G markers composed the other 13.8%. S-marker-associated genes enriched in GO terms that were related to the hallmarks of cancer, and G-markers-associated genes enriched in pathways of stem cell pluripotency and transcriptional misregulation in cancer. Significant difference in DNA methylation score was observed between peripheral normal tissue and tumor-adjacent tissues 5 mm further from the histological margin (p < 0.001 in MHB markers). DNA methylation score at and beyond 10 mm from histological margin is not significantly different from peripheral normal tissues (p > 0.05 in all markers).

Conclusions

According to the methylation pattern observed in our study, it was implied that methylation alterations were not significantly different between tissues at or beyond P10 and distal normal tissues. This finding explained for the excellent prognosis from radical resections with surgical margins of more than 15 mm. The inclusion of epigenetic characteristics into surgical margin analysis may yield a more sensitive and accurate assessment of remnant cancerous and precancerous cells in the surgical margins.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13148-021-01140-3.

Outcomes of superior segmentectomy versus lower lobectomy for superior segment Stage I non-small-cell lung cancer are equivalent: An analysis of 196 patients at a single, high volume institution

OBJECTIVES

To determine if superior segmentectomy has equivalent overall (OS), disease-free (DFS), and locoregional-recurrence-free survival (LRFS) to lower lobectomy for early-stage non-small-cell lung cancer (NSCLC) in the superior segment.

METHODS

We retrospectively reviewed all Stage 1 lower lobectomies for superior segment lesions and superior segmentectomies at our hospital from 2000 to 2018. Comparison statistics and Cox hazard modeling were performed to determine differences between groups and attempt to identify risk factors for OS, DFS, and LRFS.

RESULTS

Superior segmentectomy patients, compared with lower lobectomy patients, had more current smokers, worse forced expiratory volume in 1 s percentage, radiologic emphysema scores, clinically and pathologically smaller tumors, and more occurrences of 0 lymph nodes examined. Outcomes for superior segmentectomy compared with lower lobectomy were equivalent for 5-year OS (67.0% vs. 75.1%, p = 0.70), DFS (56.9% vs. 60.4%, p = 0.59), and LRFS (87.9% vs. 91.3%, p = 0.46). Multivariable Cox modeling lacked utility due to no outcome differences.

CONCLUSIONS

In well-selected patients, superior segmentectomies can have equivalent OS, DFS, and LRFS compared with lower lobectomies of superior segment tumors for early stage lung cancer. Further data are needed to provide better risk estimates.

Analysis of Segmental Lymph Node Metastasis and Clinical Features in cT1N0M0 Lung Adenocarcinoma

The progression of lung adenocarcinoma through lymph node metastasis has been well established; however, the process of segmental lymph node (LSN) metastasis in cT1N0M0 lung adenocarcinoma remains unclear. We aimed to elucidate the markers of lymph node metastasis to different segments in early-stage lung adenocarcinoma and identify new indications for segmentectomy. A total of 200 patients were enrolled in this study. These patients were diagnosed with cT1N0M0 lung adenocarcinoma after positron emission tomography/computed tomography and received lobectomy and lymph node dissection surgeries. Lymph nodes retrieved from each station were sorted. The metastatic status of the isolated (i) LSNs and several characteristics were analyzed. Patients with ground-glass nodules (GGNs) (P=0.025), AIS/MIA/lepidic adenocarcinoma (P=0.038), nodules with a maximum diameter ≤1 cm (P=0.017), maximum standardized uptake value (SUV_max) < 2.5 (P=0.029), serum carcinoembryonic antigen (CEA) levels ≤4.5 ng/ml (P=0.036), and no N1 lymph nodes metastasis (P=0.036) had significantly lower iLSN metastasis rates than those without these characteristics. Pure GGNs, CEA levels ≤4.5 ng/ml, SUV_max < 2.5, tumors with a maximum diameter of ≤1 cm, or those confirmed to be adenocarcinoma in situ, minimally invasive adenocarcinoma, or invasive lepidic-predominant adenocarcinoma by frozen section may indicate segmentectomy. However, segmentectomy is not suitable for patients with metastasis to the N1 lymph nodes.

Prevalence of Occult Peribronchial N1 Nodal Metastasis in Peripheral Clinical N0 Small (≤2 cm) Non-Small Cell Lung Cancer

BACKGROUND

There has been growing interest in limited resection and nonsurgical treatment for small lung cancers. We examined the pattern and rate of occult N1 nodal metastasis in patients with peripheral, small (≤2 cm), clinically node-negative non-small cell lung cancer (NSCLC).

METHODS

Patients with peripheral small (≤2 cm) NSCLC with no evidence of locally advanced or metastatic disease (clinical T1a-b N0 M0, American Joint Committee on Cancer 8th Edition Cancer Staging Manual), who were deemed eligible for lobectomy or sublobar resection, were identified from preregistration eligibility screening logs for the Alliance/Cancer and Leukemia Group B 140503 trial at our institution. Pathologic outcomes were examined in all patients undergoing anatomic resection with mediastinal and hilar lymphadenectomy.

RESULTS

Included were 58 patients treated between November 2014 and January 2017 who met the inclusion criteria: 51 underwent lobectomy, and 7 underwent segmentectomy. Mean tumor diameter on computed tomography was 1.5 cm, and mean positron emission tomography maximal standardized uptake value was 3.9. The mean consolidation-to-tumor ratio was 0.77. Occult nodal metastases in N1 stations were found in 8 of 58 patients (14%), and most of these nodes were found in interlobar or peribronchial stations (11 or 12). An additional 2 patients (3%) had occult positive N2 nodes. Overall, the false-negative rate for clinical staging was 16%.

CONCLUSIONS

Occult nodal disease was frequently identified in peripheral N1 stations (11-13) in patients with small (≤2 cm) clinical N0 NSCLC. Hilar lymphadenectomy is essential for accurate staging in the management of patients with small clinical N0 NSCLC.

Long-Term Results for Clinical Stage IA Lung Cancer: Comparing Lobectomy and Sublobar Resection

BACKGROUND

Lobectomy has been compared with sublobar resection for the treatment of stage IA non-small cell lung cancer (NSCLC). Accurate long-term data are lacking on the risk of recurrence in routine clinical practice. This study used a unique and representative dataset to compare recurrence, overall survival (OS), and lymph node staging between lobectomy and sublobar resection.

METHODS

The American College of Surgeons performed a Special Study of the National Cancer Data Base, by reabstracting records to augment NSCLC data with enhanced information on preoperative comorbidity and cancer recurrence from 2007 to 2012. For patients treated with lobectomy or sublobar resection (wedge resection or segmentectomy) for clinical stage IA NSCLC, propensity matching and competing risks models compared 5-year OS and risk of cancer recurrence. Secondary measures included lymph nodes collected, pathologic upstaging, and surgical margin status.

RESULTS

A total of 1,687 patients with stage IA NSCLC were identified (1,354 who underwent lobectomy, and 333 who had sublobar resections). Propensity matching yielded 325 pairs. Lobectomy and sublobar resection groups had similar 5-year OS (61.8% vs 55.6%, p = 0.561). The sublobar group had a 39% increased risk of NSCLC recurrence (hazard ratio, 1.39; 95% confidence interval, 1.04 to 1.87). Median lymph node counts were higher for lobectomy-treated patients (7 [3, 10] vs 1 [0, 4], p < 0.001)].

CONCLUSIONS

In an enhanced national dataset representative of outcomes for stage IA NSCLC, sublobar resection was associated with a 39% increased risk of cancer recurrence. The majority of patients treated with sublobar resection had an inadequate lymph node assessment. These real-world results must be considered when existing clinical trial results comparing these treatments are extrapolated for clinical use.