Postoperative surveillance for renal cell carcinoma: a multifactorial histological subtype specific protocol

OBJECTIVE To create a model that adjusts surveillance after surgery to the natural history of surgically treated renal cell carcinoma (RCC), and to assess the cost of several surveillance models with a long-term longitudinal follow-up, as although there are many models for predicting the outcome in RCC, most surveillance protocols remain based primarily on stage alone, and thus might be inaccurate as they do not incorporate many other pathological features that have a significant effect on recurrence. PATIENTS AND METHODS We identified 1864, 357 and 118 patients with pM0 clear cell, papillary and chromophobe RCC, respectively, who had a a radical or partial nephrectomy between 1970 and 2000. All recurrences were classified according to location (abdomen, thorax, bone, brain). Cox proportional hazards models were used to determine which pathological features were independently predictive of recurrence in each group. Three subtype-specific protocols were devised based on site-specific recurrence rates. RESULTS Positive surgical margins, the 2002 Tumour-Node-Metastasis classification, size, nuclear grade, and histological tumour necrosis were independently associated with abdominal recurrence in patients with clear-cell RCC. These same features, except for surgical margins, were significantly associated with thoracic recurrence. The 2002 classification and nuclear grade were independently associated with abdominal and thoracic recurrence in patients with papillary RCC. No multivariate analysis was done for chromophobe RCC as there were only 10 recurrences to the abdomen and three to the thoracic region. However, these patients were stratified according to stage and grade, as recurrences in this group had a clear stage- and grade-specific pattern. CONCLUSIONS We present a subtype-specific multifactorial surveillance protocol based on significant predictors of recurrence. This protocol is better than algorithms based on stage alone and can be used to effectively tailor postoperative imaging to the individual patient.