Optical biomarker analysis for renal cell carcinoma obtained from preoperative and postoperative patients using...

Renal cell carcinoma (RCC) is the most common malignant tumor in the urinary system, accounting for 80 % to 90 % for all renal malignancies. Traditional diagnostic methods like magnetic resonance imaging (MRI) and computed tomography (CT) lack the sensitivity and specificity as they lack specific biomarkers. These limitations impede effective monitoring of tumor recurrence. This study aims to employ Attenuated Total Reflection (ATR)-Fourier transform infrared (FTIR) spectroscopy, an optical technology sensitive to molecular groups, to analyze the potential optical biomarkers in urine and plasma samples from RCC patients pre- and post-surgery. The results reveal distinctive spectral information from both plasma and urine samples. Post-surgery urine spectra exhibit complexity compared to plasma, showing reduced content at 1072 cm^-1, 1347 cm^-1 and 1654 cm^-1 bands, while increased content at 1112 cm^-1, 1143 cm^-1, 1447 cm^-1, 3334 cm^-1 and 3420 cm^-1 bands. Utilizing machine learning models such as eXtreme Gradient Boosting (XGBoost), support vector machine (SVM), partial least squares (PLS), and artificial neural network (ANN), the study evaluated plasma and urine samples pre- and post-surgery. Remarkably, the XGBoost method excelled in distinguishing between tumor conditions and recovery, achieving an impressive AUC value of 0.99. These results underscore the potential of ATR-FTIR technology in identifying RCC optical biomarkers, with XGBoost showing promise as a valuable screening tool for RCC recurrence diagnosis.