A proposal for a TNM staging system for extramammary Paget disease: Retrospective analysis of 301 patients with invasive primary tumors

• TNM staging system for iEMPD has not yet been established.
• To establish TNM staging system, we retrospectively collected 301 patients with invasive extramammary Paget’s disease.
• 51 had distant metastasis and 112 had LN metastasis at diagnosis.
• Distant metastasis, 2 or more lymph node metastases, lymphovascular invasion and tumor thickness over 4 mm correlated with worse survival.
• We propose a TNM staging system for EMPD using simple factors for classification (distant metastasis, LN metastasis and local tumor status).
• Each stage had a statistically distinct survival curve.

BackgroundAlthough extramammary Paget disease (EMPD) usually appears as carcinoma in situ, it sometimes becomes invasive (iEMPD) and fatal. However, a TNM staging system for iEMPD has yet to be established.ObjectiveThe aim of this study was to establish a TNM staging system for iEMPD.MethodsWe retrospectively collected iEMPD patients treated at 12 institutes in Japan. Factors reported to be associated with survival such as distant metastasis, lymph node (LN) metastasis, and primary tumor status were evaluated using the log-rank test.ResultsWe enrolled 301 iEMPD patients, of whom 114 had remote metastases (49 had both distant and LN metastasis; 2, distant metastasis only; and 63, LN metastasis only) and the remaining 187 patients had no remote metastasis. Distant metastasis (M1) showed worse survival (P < 0.00001). In the analysis of the 250 patients without distant metastasis, LN metastasis also showed worse survival (P < 0.00001). Among the patients with LN metastasis, 2 or more LN metastases (N2) showed worse survival than did single LN metastasis (N1, P = 0.02). Lastly, in the analysis of the 187 patients without metastasis, tumor thickness of over 4 mm or lymphovascular invasion showed worse survival (T2, P < 0.05 and P < 0.001, respectively). Patients with neither of these features were defined as T1. From these results, we propose this TNM staging system: stage I, T1N0M0; stage II, T2N0M0; stage IIIa, anyTN1M0; stage IIIb, anyTN2M0; stage IV, anyTanyNM1. Other than stages II and IIIa, each stage had a statistically distinct survival curve.ConclusionWe propose a TNM staging system for EMPD using simple factors for classification that could provide important prognostic information in managing EMPD. However, accumulation of more patient data and further revision of the system are required.