Introduction

In absence of objective criteria for staging bone scintigraphy [BS], the decision to perform preoperative BS for patients diagnosed with renal cell carcinoma [RCC] remains a subjective practice. We aimed at identifying an objective and reproducible strategy for preoperative bone staging in RCC patients.

Methods

We evaluated 2,008 RCC patients treated with surgery and prospectively included into an institutional database. The study outcome was the presence of one or more lesions suspicious for bone metastases at staging BS. In case of uncertain result, BS was considered negative in case of no-evidence of progression during follow-up or positive in all other cases. Patients without a pre-operative BS but with a negative post-operative BS were considered negative by definition. A multivariable logistic regression model was fitted to predict positive BS. Predictors consisted of preoperative platelet/haemoglobin [PLT/HB] ratio, clinical tumour stage [cT], clinical nodal stage [cN] and presence of systemic symptoms. A 2000-sample bootstrap validation was used to estimate H-index. Decision curve analysis [DCA] was used to compare the performance of the proposed model with the prediction based on symptoms only, as recommended by guidelines.

Results

BS resulted negative in 1927 (96%) patients and positive in 81 patients (4%). Preoperative PLT/HB ratio was associated with higher risk of positive BS (Odds Ratio [OR] 1.04; p<0.001). Similarly, cT (cT2 vs cT1a; OR 2.13; p=0.02), cN (cN1 vs cN0; OR 2.5; p=0.001) and presence of systemic symptoms (OR 4.26; p<0.001) were all associated with higher risk of positive BS. Following a 2000-sample bootstrap validation, H-index of the proposed model was 0.76 whereas the H-index of the prediction based on systemic symptoms only resulted 0.63. At DCA, the net benefit of the proposed model was superior than the net benefit of the prediction based on symptoms only.

Conclusions

Based on the proposed model, it is possible to accurately estimate the risk of positive BS at kidney cancer staging using pre-operative characteristics. If BS is performed only when the risk of positive result is >5%, a negative BS is spared in 80% of the population and a positive BS is missed in 2% of the population only. When compared to decision-making based on symptoms only, which represents the strategy recommended by available guidelines, the proposed model resulted more objective, statistically more accurate and clinically associated with higher net benefit. These figures support an update of the available guidelines.

Funding

None