Prognostic role of the log odds of positive lymph nodes in Western patients with resected gastric cancer: A comparison with the 8th edition of the TNM staging system

OBJECTIVES

Several alternative lymph node staging systems have recently been described for gastric cancer. The log odds of positive lymph nodes (LODDS) system may be superior to the pN stage (American Joint Committee on Cancer) and lymph node ratio systems in predicting outcomes for patients with gastric cancers, as indicated by some researchers. Most studies, however, have been conducted in Asian countries, and conflicting results have been reported by other investigators.

METHODS

We conducted a retrospective study of all 377 cases of gastric cancer resected at a tertiary hospital in Spain between 2000 and 2019. Clinicopathologic features were collected, LODDS were calculated and categorized into 5 groups (S1-S5), and statistical analysis was performed.

RESULTS

The cases included (n = 315) were classified as S1 (25.6%), S2 (18.4%), S3 (21.3%), S4 (20.3%), and S5 (14.4%). The LODDS classification was significantly associated with tumor size, Laurén subtype, presence of signet ring cells, tumor grade, perineural infiltration, lymphovascular invasion, growth pattern, pT, tumor recurrence, and death. Kaplan-Meier analysis based on the LODDS classification demonstrated improved patient stratification compared with the pN stage for both overall survival (OS) and disease-free survival (DFS). Area under the curve values for recurrence and death were superior for the LODDS classification, and this classification was independently related to OS and DFS. In addition, the LODDS classification successfully divided patients without lymph node metastases (pN0) into subgroups with distinct prognoses.

CONCLUSIONS

For our cohort, the LODDS system showed better prognostic performance than pN stage; it was an independent predictor of OS and DFS, and it provided valuable prognostic information in cases without lymph node metastases. Its prognostic accuracy, however, decreased in cases with fewer than 16 lymph nodes resected.

Analysis of the Effects of Melanin Depigmentation on Immunohistochemical Staining

Heavily pigmented lesions are difficult to evaluate histologically, as melanin obscures cellular details. Several classic laboratory techniques aim to clear melanin and allow evaluation. Most of them are old and appeared before immunohistochemistry (IHC) use. Many laboratories perform IHC with aminoethylcarbazole instead of diaminobenzidine (DAB) in heavily pigmented lesions, as red-stained is easy to interpret despite pigmentation. Nevertheless, many laboratories lack alternatives to DAB. The aim of this study is to compare 6 different tissue bleaching techniques and evaluate which is the best for immunohistochemical staining with DAB. In the present study we have selected a case with gross pigmentation because of the high grade of melanin deposition. We have performed 6 different bleaching techniques and subsequently performed 2 different IHC stains, frequently used in melanoma: SOX10 (nuclear) and Melan-A (cytoplasmic). Five different pathologists, 2 of them with expertise in dermatopathology, have blindly reviewed and scored the staining quality. Our results indicate a high grade of interobserver concordance in the evaluation of IHC results between pathologists. All the bleaching techniques that included a sulfuric acid led to tissue detachment from the slide. The best method for SOX10 was that based in potassium permanganate, with a high quality of staining (4 over 5), while the best method for Melan-A was the 1 based in peroxide hydrogen (4 over 5). We consider this study can be quite useful for those laboratories lacking aminoethylcarbazole for IHC techniques, allowing the use of DAB for IHC of heavily pigmented lesions.

Towards standardization of lymph-node ratio classifications: Validation and comparison of different lymph node ratio classifications for predicting prognosis of patients with resected gastric cancer

INTRODUCTION

The TNM staging system is the main prognostic tool for GC, but the number of metastatic lymph nodes (LN) can be affected by surgical, pathological, tumor or host factors. Several authors have shown that lymph node ratio (LNR) may be superior to TNM staging in GC. However, cut-off values vary between studies and LNR assessment is not standardized.

MATERIAL AND METHODS

Retrospective study of all GC resected in a western tertiary center (N = 377). Clinical features were collected and pathological features were assessed by two independent pathologists. Eight LNR classifications were selected and applied to our patients. Statistical analyses were performed.

RESULTS

315 patients were included. Most tumors were T3 (49.2%) N+ (59.3%). During follow-up, 36.7% of patients progressed and 27.4% died due to tumor. All LNR classifications were significantly associated with clinicopathological features such as Laurén subtype, lymphovascular invasion, perineural infiltration, T stage, tumor progression or death. All LNR classifications were independent prognostic factors for OS and DFS, and ROC analyses calculated similar AUC values for all staging systems. Kaplan-Meier curves showed that Pedrazzani, Wang, Liu and Huang classifications stratified patients better into three (Pedrazzani) or four categories. These classifications tended to downstage TNM N2 and N3 tumors. In cases with less than 16 LNs resected, Pedrazzani and Wang classifications showed the best prognostic performance.

CONCLUSIONS

Pedrazzani, Wang, Liu and Huang classifications showed good prognostic performance in western GC patients. Larger studies in other cohorts are needed to identify the most consistent LNR classification for GC.

Development of a simplified tumor-lymph node ratio classification system for patients with resected gastric cancer: A western study

INTRODUCTION

Gastric cancer (GC) shows high recurrence and mortality rates. The AJCC TNM staging system is the best prognostic predictor, but lymph node assessment is a major source of controversy. Recent studies have found that lymph node ratio (LNR) may overcome TNM limitations. Our aim is to develop a simplified tumor-LNR (T-LNR) classification for predicting prognosis of resected GC.

METHODS

Retrospective study of all GC resected in a tertiary center in Spain (N = 377). Clinicopathological features were assessed, LNR was classified into N0:0%, N1:1-25%, N2:>25%, and a T-LNR classification was developed. Statistical analyses were performed.

RESULTS

317 patients were finally included. Most patients were male (54.6%) and mean age was 72 years. Tumors were intestinal (61%), diffuse (30.8%) or mixed (8.1%). During follow-up, 36.7% and 27.4% of patients progressed and died, respectively. T-LNR classification divided patients into five prognostic categories (S1-S5). Most cases were S1-S4 (26.2%, 19.9%, 22.6% and 23.6%, respectively). 7.6% of tumors were S5. T-LNR classification was significantly associated with tumor size, depth, macroscopical type, Laurén subtype, signet ring cells, histologic grade, lymphovascular invasion, perineural infiltration, infiltrative growth, patient progression and death. Kaplan-Meier curves for OS showed an excellent patient stratification with evenly spaced curves. As for DFS, T-LNR classification also showed good discriminatory ability with non-overlapping curves. T-LNR classification was independently related to both OS and DFS.

CONCLUSIONS

T-LNR classifications can successfully predict prognosis of GC patients. Larger studies in other geographic regions should be performed to refine this classification and to validate its prognostic relevance.