Nuclear morphometry accurately predicts recurrence in clinically localized renal cell carcinoma

Despite careful clinical staging, as many as 30 percent of patients with pathologically, specimen-confined renal cell carcinoma (RCC) have unpredictable recurrence following surgery. Present pathologic and clinical staging systems cannot accurately predict those patients with high risk of disease recurrence from those who are cured by surgery alone. Advances in immunotherapy including gene therapy for RCC have dictated the need to identify RCC patients for adjuvant therapy protocols who have a high probability of recurrence following nephrectomy. Nuclear morphometric techniques developed at our institution have predicted prognosis for a variety of genitourinary tumors; it was used to predict recurrence among patients undergoing nephrectomy for localized RCC. This report is a retrospective study of 26 patients with RCC of similar age, stage (pT1-pT3), and grade. Fifteen were free of disease at a mean of 75.2 months, and 11 had distant disease recurrence at a mean of 27.1 months. Statistical analysis of a variety of nuclear shape descriptors accurately separated this group of patients based on disease recurrence. No nuclear shape descriptor predicted disease recurrence when nuclei within the region of the tumor with the highest grade were analyzed. However, the range of nuclear ellipticity (p = 0.007) best predicted disease recurrence when nuclei were selected in a random fashion. Multivariate analysis of the four best shape descriptors better predicted disease recurrence (p = 0.002) with a sensitivity of 73 percent and specificity of 100 percent. These results are encouraging and suggest that this technique might be used in identifying patients for adjuvant gene therapy.