The Performance of a Dual-Energy CT Derived Radiomics Model in Differentiating Serosal Invasion for Advanced Gastric Cancer Patients After Neoadjuvant Chemotherapy: Iodine Map Combined With 120-kV Equivalent Mixed Images

Objectives

The aim was to determine whether the dual-energy CT radiomics model derived from an iodine map (IM) has incremental diagnostic value for the model based on 120-kV equivalent mixed images (120 kVp) in preoperative restaging of serosal invasion with locally advanced gastric cancer (LAGC) after neoadjuvant chemotherapy (NAC).

Methods

A total of 155 patients (110 in the training cohort and 45 in the testing cohort) with LAGC who had standard NAC before surgery were retrospectively enrolled. All CT images were analyzed by two radiologists for manual classification. Volumes of interests (VOIs) were delineated semi-automatically, and 1,226 radiomics features were extracted from every segmented lesion in both IM and 120 kVp images, respectively. Spearman's correlation analysis and the least absolute shrinkage and selection operator (LASSO) penalized logistic regression were implemented for filtering unstable and redundant features and screening out vital features. Two predictive models (120 kVp and IM-120 kVp) based on 120 kVp selected features only and 120 kVp combined with IM selected features were established by multivariate logistic regression analysis. We then build a combination model (ComModel) developed with IM-120 kVp signature and ycT. The performance of these three models and manual classification were evaluated and compared.

Result

Three radiomics models showed great predictive accuracy and performance in both the training and testing cohorts (ComModel: AUC: training, 0.953, testing, 0.914; IM-120 kVp: AUC: training, 0.953, testing, 0.879; 120 kVp: AUC: training, 0.940, testing, 0.831). All these models showed higher diagnostic accuracy (ComModel: 88.9%, IM-120 kVp: 84.4%, 120 kVp: 80.0%) than manual classification (68.9%) in the testing group. ComModel and IM-120 kVp model had better performances than manual classification both in the training (both p<0.001) and testing cohorts (p<0.001 and p=0.034, respectively).

Conclusions

Dual-energy CT-based radiomics models demonstrated convincible diagnostic performance in differentiating serosal invasion in preoperative restaging for LAGC. The radiomics features derived from IM showed great potential for improving the diagnostic capability.

Radiologic serosal invasion sign as a new criterion of T4a gastric cancer on computed tomography: diagnostic performance and prognostic significance in patients with advanced gastric cancer

PURPOSE

To investigate the diagnostic performance and prognostic significance of a new criterion for radiologic T4a staging on computed tomography (CT) in patients with advanced gastric cancer (AGC).

METHODS

Between January 2010 and April 2019, 101 patients with pathologically confirmed gastric cancer were collected. Among them, 53 patients with pathologic T3 and T4a cancers were included in this study. Three reviewers assessed preoperative CT scans for radiologic T staging in two sessions, independently and in consensus at a 2-week interval, while blinded about the pathologic T stage. The radiologic serosal invasion sign was defined as a nodular extension from the outer gastric wall reaching beyond the perigastric vascular plane and adopted as a new CT criterion for T4a cancer. We evaluated the diagnostic performance, interobserver agreement, and prognostic significance of this sign for the postoperative recurrence.

RESULTS

There were 46 pathologic T3 cancers (86.7%) and seven pathologic T4a cancers (13.2%). The diagnostic performance of the radiologic serosal invasion sign in the differentiation between T3 and T4a cancers was as follows: sensitivity, 91.3%; specificity, 71.43%; and accuracy, 88.68% for R1 and sensitivity, 78.26%; specificity, 85.71%; and accuracy, 79.25% for R2. The k-value was 0.64. Among the clinical and pathologic variables, radiologic T4a sign [hazard ratio (HR): 7.96; 95% confidence interval (CI) 2.36-26.86, p = 0.001], pathologic T4a (HR 9.82, 95% CI 2.35-40.95, p = 0.002), tumor size (HR 1.18, 95% CI 1.02-1.35, p = 0.026), and lymphovascular invasion (HR 6.39, 95% CI 1.42-28.75, p = 0.015) were the significant factors for postoperative recurrence.

CONCLUSIONS

Radiologic serosal invasion sign is reliable as a new CT criterion for T4a cancer staging in patients with advanced gastric cancer, demonstrating 80% to 88% accuracy. Radiologic serosal invasion sign can also serve as a prognostic factor for postoperative recurrence as well as pathologic T4a stage.

Clinical and pathological staging of gastric cancer: Current perspectives and implications

Accurate categorization of invasive depth and lymph node metastasis or optimization of TNM categories is fundamentally critical for prognostic assessment and decision making regarding subsequent therapies after surgery for gastric cancer. Improving the precision of the TNM staging is the ongoing goal. The evolution of the staging system indicates that there is no "ideal staging". Every update has criticized the lack of a standard approach for the stages to date. T staging depends on the accurate determination of the depth of infiltration based on pathological continuous sections. N staging is susceptible to the influence of lymph node detection, and insufficient lymph node detection can lead to N staging migration. M staging is required to improve the detection rate of peritoneal positive free cancer cells to determine the high risk factors of peritoneal metastasis. At present, the quality of standardized pathological diagnosis of gastric cancer requires improvement. Based on a review of the literature and experience from multiple gastric cancer centers, we present a new development in TNM staging and a way to improve clinical and pathological quality control of gastric cancer.

Improvement of T stage precision by integration of surgical and pathological staging in radically resected stage pT3-pT4b gastric cancer

Background

Both surgical TNM (sTNM) and pathological TNM (pTNM) staging are important clinicopathologic indexes of gastric cancer (GC). However, surgeons and pathologists might assess tumor depth differently in the same patient. To investigate the prognostic significance of sTNM status in patients with radically resected stage pT3-pT4b GC, we examined the relationship between sTNM and pTNM.

Methods

Clinicopathologic and survival data of 1289 patients with stage pT3-pT4b GC were studied retrospectively, in the aftermath of radical surgery.

Results

The unconformity for assessing tumor invasion depth were frequently exhibited between sT and pT staging. Comparison of 5-year OS among them, no significant differences were observed (pT3/sT3 vs pT3/sT4a, p=0.962; pT4a/sT4b vs pT4b/sT4b, p=0.508). Also, pT3/sT4b, pT4a/sT3 and pT4a/sT4a were homogeneity in prognosis. We proposed a revised pT stage in which surgical macroscopic T4b (sT4b) was incorporated into the pT stage, namely, patients in the pT3 stage with sT4b cancers were reclassified as being in the r-pT4a stage; patients in the pT4a stage with sT4b cancers were reclassified as being in the r-pT4b stage. In two-step multivariate analysis, revised pT stage proved more suitable for determining prognosis, surpassing both UICC/AJCC pT stage and sT stage as an independent prognostic index.

Conclusions

Surgical T stage is a significant and independent prognostic index of overall survival (OS) in patients with radically resected advanced GC. Patients in the pT3/4a stage with sT4b cancers, are potentially underestimated, and should be considered higher stage in terms of prognostic.

Prognostic significance of intraoperative macroscopic serosal invasion finding when it shows a discrepancy in pathologic result gastric cancer

Purpose

Depth of wall invasion is an important prognostic factor in patients with gastric cancer, whereas the prognostic significance of intraoperative macroscopic serosal invasion (mSE) findings remain unclear when they show a discrepancy in pathologic findings. This study, therefore, assessed the prognostic significance of mSE.

Methods

Data from cohort of 2,835 patients with resectable gastric cancer who underwent surgery between 1990 and 2010 were retrospectively reviewed.

Results

The overall accuracy of mSE and pathologic results was 83.4%. The accuracy of mSE was 75.5% in pT2. On the other hand, the accuracy of pT3 dropped to 24.5%. According to mSE findings (+/–), the 5-year disease-specific survival (DSS) rate differed significantly in patients with pT2 (+; 74.2% vs. –; 92.0%), pT3 (+; 76.7% vs. –; 91.8%) and pT4a (+; 51.3% vs. –; 72.8%) (P < 0.001 each), but not in patients with T1 tumor. Multivariate analysis showed that mSE findings (hazard ratio [HR], 2.275; 95% confidence interval [CI], 1.148–4.509), tumor depth (HR, 6.894; 95% CI, 2.325–20.437), nodal status (HR, 5.206; 95% CI, 2.298–11.791), distant metastasis (HR, 2.881; 95% CI, 1.388–6.209), radical resection (HR, 2.002; 95% CI, 1.017–3.940), and lymphatic invasion (HR, 2.713; 95% CI, 1.424–5.167) were independent predictors of 5-year DSS rate.

Conclusion

We observed considerable discrepancies between macroscopic and pathologic diagnosis of serosal invasion. However, macroscopic diagnosis of serosal invasion was independently prognostic of 5-year DSS. It suggests that because the pathologic results could not be perfect and the local inflammatory change with mSE(+) could affect survival, a combination of mSE(+/–) and pathologic depth may be predictive of prognosis in patients with gastric cancer.

Borrmann type IV gastric cancer should be classified as pT4b disease. J Surg Res. 2016;203:258-267. [PMID: 27363630 DOI: 10.1016/j.jss.2016.04.026]

BACKGROUND

The impact of macroscopic pathologic features of primary tumor that could be obtained preoperatively on pT classification has not been reported so far. The aim of this study was to investigate the feasibility of incorporation of Borrmann type IV gastric cancer into the pT classification.

MATERIALS AND METHODS

Clinicopathologic and prognostic data of 1622 patients with advanced gastric cancer who underwent radical surgery were retrospectively studied.

RESULTS

Of 1622 patients, 135 (8.32%) patients were classified as having Borrmann type IV gastric cancer. We first confirmed that Borrmann type IV gastric cancer was one of the independent prognostic factors for patients with advanced gastric cancer who underwent radical surgery. Interestingly, we found that overall survival of patients with Borrmann type IV gastric cancer could be clearly distinguished by pN classification and pathological TNM stage but not by pT classification. Importantly, further analysis demonstrated that the prognosis of Borrmann type IV gastric cancers was homogeneous with that of pT4b cancers but poorer than pT2, pT3, pT4a cancers. Therefore, we proposed a novel pT classification in which pT4b disease was defined as cancers that were Borrmann type IV or those that had invaded adjacent structures. Two-step multivariate analysis demonstrated that the novel pT classification was more suitable for prognostic assessment than the original classification.

CONCLUSIONS

Classifying Borrmann type IV gastric cancer as pT4b disease improves pT classification prediction of prognosis in patients with advanced gastric cancer after radical surgery.