br Statistical analyses were performed using SPSS
Statistical analyses were performed using SPSS (version 23; IBM, Armonk, N.Y., U.S.A.) or R version 3.3.0 (R Foundation for Statistical Computing, Vienna, Austria). Mean values, median, interquartile ranges, and/or ranges are presented for quantitative data as appropriate. Absolute and relative frequencies are given for categorical data. Performance of ex vivo radioactive rating (positive vs negative) for resected tissue specimens is described using sensitivity, specificity, and accuracy.
For estimation of sensitivities and corresponding confidence intervals, an intercept-only logistic generalized estimating equation (GEE) model accounting for multiple measurements in one patient was fit to the data. The result of the dichotomized test was used as a dependent variable, and only patients with a positive histopathological result were considered .
To derive estimates for the specificities, a variable indicating whether a negative test result was observed was used as a dependent variable. Here, only patients with a negative histopathological result were included. Accuracy was estimated in an intercept-only model with a dependent variable that L-Glutamine indicated whether the test result and the result of the histopathological assessment agreed. For all GEE models, an independent correlation structure was assumed. Waterfall plots were used to show the best PSA responses after 99mTc-PSMA-RGS. The distribution of biochemical-free survival and PC-specific treatment–free survival times after 99mTc-PSMA-RGS was estimated using the Kaplan-Meier method.
3.1. 99mTc-PSMA-RGS and comparison of gamma probe measurements with histopathology
99mTc-PSMA-RGS was able to identify and remove all lesions detected on preoperative 68Ga-PSMA-11 PET in all patients. In total, 132 surgical specimens were removed (median specimens per patient: 4; range: 1–10). Forty-six specimens were classified as positive and 86 were considered negative.
The count rate of the background reference ranged between and 4 counts/s. For positive specimens, the median count rate during ex vivo gamma probe measure-ments was 21.5 (range: 4–246) counts/s, while negative specimens showed a median count rate of 0.5 (range: 0–4) counts/s.
Fig. 1 – Overview of steps for PSMA-radioguided surgery: step 1: selection of patients based on PSMA PET results and clinical history; step 2: injection of 99mTc-PSMA-I&S and subsequent SPECT/CT imaging to confirm 99mTc-PSMA-I&S uptake in same lesions of preoperative 68Ga-PSMA-11 findings; step
3: PSMA-radioguided surgery is performed with in vivo and ex vivo gamma probe measurements to reliably identify metastatic prostate cancer lesions. CT = computed tomography; 68Ga-PSMA-11 = 68Ga-PSMA N,N0 -bis[2-hydroxy-5-(carboxyethyl)benzyl] ethylenediamine-N,N0 -diacetic acid; PET = positron emission tomography; PSMA = prostate-specific membrane antigen; SPECT = single photon emission computed tomography; 99mTc-PSMA-I&S = 99mTc-PSMA investigation and surgery.
In the seven patients with false-negative findings, histopathology revealed a total of 12 metastatic lesions that were detected neither on preoperative 68Ga-PSMA-11 PET nor during 99mTc-PSMA-RGS. In one of these patients harboring three undetected lesions (maximum diameter 5 mm), microspores might be attributed to a low amount of tracer injected (221 MBq, 2.3 MBq/kg) and a long time interval between injection and surgery (22.2 h). In the other six
patients, lesions had a median size of 2 mm (range: 1– 4 mm). The median size of correctly identified metastatic lesions during 99mTc-PSMA-RGS was 12 mm (range: 3– 25 mm). Moreover, compared with preoperative 68Ga-PSMA-11 PET, 99mTc-PSMA-RGS detected additional metas-tases as small as 3 mm in two patients. Of note, due to inherent limitations of SPECT/CT, imaging with 99mTc-PSMA-I&S was able to detect only 25 (56.8%) of the 44 lesions observed on 68Ga-PSMA-11 PET.