Objective Tumor sink effect (TSE) has been defined as; decreased uptake in healthy tissue with increased tumor sequestration of the radiopharmaceuticals. It enables us to give high tumoral radiation doses while resulting in lower absorbed radiation to critical organs. However, the factors which influence this effect are yet to be defined. In this study, we have investigated the predictive factors of the tumor sink effect in a group of patients who received Lu-177-Prostate-specific membrane antigen (PSMA) therapy due to progressive metastatic castration-resistant prostate cancer (mCRPC). Methods We have retrospectively analyzed the pre-therapy Ga-68-PSMA positron-emission tomography (PET)-computed tomography (CT) and post-therapy planar whole-body scans of 65 patients who received at least two cycles of 7.4 GBq of Lu-177-PSMA therapy. All patients with mCRPC were referred to our department after multiple treatment lines. Age, previous therapies, International Society of Urological Pathology (ISUP) score, and pre-therapy serum tumor marker levels were recorded. Post Lu-177-PSMA therapy images were analyzed for TSE. Ga-68-PSMA PET-CT images were used for the calculation of SUVmax in malignant and healthy tissues as well as metabolic tumor volume (MTV) and total lesion PSMA index (TLPI). Results Based on the post-therapy scans, TSE was seen in 17/65 (26.2%) patients. In univariate analysis, patients with TSE had higher pre-therapy PSA, PSA velocity, and ALP (p < 0.0001). In relation to PET parameters, patients with TSE had higher Ga-68-PSMA MTV, Ga-68-PSMA TLPI and lower pretherapy renal SUVmax (p < 0.0001)), pretherapy liver SUVmax (p:0.012), pretherapy parotid gland SUVmax (p:0.032), and pretherapy parotid gland SUVmean (p:0.038). In the multivariant analysis, Ga-68-PSMA TLPI, pre-therapy PSA, and PSA velocity were found to be statistically significant. When analyzed according to Youden index, pretherapy PSA level of 133 ng/ml (sensitivity 0.765 and 0.875), PSA velocity of 246 ng/ml/year (sensitivity 0.765 and 0.833), and Ga-68-PSMA TLPI of 2969 g (sensitivity 0.765 and 0.875) was found to be the best cut-off points to predict TSE. Conclusion The tumor sink effect was seen in 26.2% of patients. Ga-68- PSMA TLPI, pre-therapy PSA, and PSA velocity was found to be the predictors of TSE. Accurate prediction of TSE may lead to increased tumoral doses while sparing healthy organs. Clinical trials that consider this effect as a part of a dose algorithm may further increase therapeutic efficacy.