Patterns of Discordance Between Clinical and Pathologic Stage in Head and Neck Squamous Cell Carcinoma

N. Punjabi,B. Hondorp,D. Macias,Y. Liu,J. Inman

DOI: https://doi.org/10.1016/j.ijrobp.2024.01.151

IF: 8.013

2024-04-01

International Journal of Radiation Oncology*Biology*Physics

Abstract:Purpose/Objective(s) Like other malignancies, treatment and prognosis of head and neck squamous cell carcinoma is highly dependent on TNM staging. Clinical TN staging (cTN), which is established during the primary workup, helps guide initial treatment but is limited by the quality of physical examination and imaging studies. Pathologic TN staging (pTN) is determined from collected specimens if surgical resection occurs and is considered the gold standard for staging. Previous studies have shown discordance between cTN and pTN, likely due to the limitations of clinical staging. The aim of this study is to define patterns of discordance between clinical and pathological T staging in multiple subsites of head and neck cancer. Materials/Methods A retrospective cohort of 580 surgically treated head and neck squamous cell carcinoma patients from a single institution is presented. cTN and pTN were compared in the overall cohort and at specific anatomical subsites, including the oral cavity, oropharynx, and larynx. Cohen's kappa (κ) statistic was used to assess the “inter-rater reliability” between cTN and pTN. Results Nearly 1/3 of cases had staging discordance. Cohen's kappa coefficient for the entire cohort was κ =0.55, indicating moderate agreement. Highly discordant stages/subsites with κ <0.45 included T2 oral cavity, T2 oropharynx, T3 larynx, and N1 neck. Oral cavity lesions T2-4 were often overstaged, and more than 1/3 of T3 larynx cancers were understaged. Highly concordant stages/subsites with κ >0.65 included T1 larynx, T4 oropharynx, N0 neck, and N3 neck.In general, there was higher concordance between cTN and pTN at the extremes of the staging systems (i.e. T1/N0 and T4/N3) while intermediate stages had higher discordance. Distribution of our cohort across each stage and Cohen's kappa coefficient for overall T and N stage are presented in the table. Groups with concordant clinical and pathologic stages are highlighted in blue. Highly discordant κ values (<0.45) are indicated by an asterisk, while highly concordant values (>0.65) are indicated by two asterisks. Conclusion Our results provide data on specific TN stages and anatomic sites where discordance is more likely to occur. Patients with discordant staging may initially receive inadequate counseling or inappropriate treatment. For example, since T3 larynx cancer was often understaged, less aggressive treatment may incorrectly be selected in this setting. The opposite may be true in T2-4 oral cancer, which was often overstaged. Further investigation is warranted to improve clinical staging accuracy in the areas of highest discordance.

oncology,radiology, nuclear medicine & medical imaging