Purpose: To explore an optimized grade group (oGG) criterion from systematic biopsies (SB) and targeted biopsies (TB) and offer a better prediction of radical prostatectomy (RP) grade group (GG).Methods: Positive needles were collected from 146 patients who underwent SB + TB followed by RP. The grade was assigned for two different kinds of biopsies with five GG criteria: (1) global GG (gGG); (2) most common GG (most common GG from SB + TB, mGG); (3) highest GG (highest numerical GG from SB + TB, hGG); (4) largest volume/linear length cancer GG (defined as GG from the SB + TB with the largest length of cancer in a needle, lGG). These biopsy grades were compared (equivalence, upgrade, or downgrade) with the final grade of the RP lesion, using weighted κ coefficients; (5) Then the best agreement of the (2) (3) (4) grading scores from SB or TB was combined to introduce an oGG.Results: In this study, gGG showed generally poor agreement (47.2%) with RP GG (weighted κ: 0.43). Using the three criteria (mGG, hGG, and lGG) of SB, mGG had the best agreement (55.5%, weighted κ: 0.46), while hGG and lGG had a lower agreement (48.6% and 48.6%, weighted κ: 0.42 and 0.38). Using the three criteria (mGG, hGG and lGG) of TB: lGG had the best agreement (56.8%, weighted κ: 0.43), while mGG and hGG had lower agreement (50.0% and 49.3%, weighted κ: 0.40 and 0.40); Then oGG was generated (higher GG between mGG of SB and lGG of TB) and the agreement of oGG increased to 59.6% and weighted κ was 0.49. Additionally, oGG had a lower upgrade rate than gGG, while the downgrade rate remained unchanged.Conclusions: oGG showed better agreement with RP GG than gGG. oGG had a lower upgrade rate than gGG, while downgrade rate remained unchanged.