Abstract Study question Does the standard, not-dilated seminiferous tubules (STs) diameter vary according with different testis histology patterns in patients with non-obstructive azoospermia (NOA)? Summary answer The standard STs diameter differed significantly among cases with different testis histology: hypospermatogenesis (HYPO) had the highest STs diameter compared to the other histological subgroups. What is known already During microdissection testicular sperm extraction (mTESE), the identification of dilated STs, as subjectively evaluated at high magnification by comparing their apparent caliber with that of the surroundings, is crucial to identify residual foci of intact spermatogenesis and to retrieve sperm. Previous studies have demonstrated that dilated STs contain sperm in most cases, but it is not clear why in some cases an apparent normal tubular caliber does not correspond to spermatogenesis integrity. Aim of the present study was to assess whether different histology patterns could affect the STs diameter. Study design, size, duration We retrospectively evaluated 168 patients with NOA undergoing unilateral (N = 91) or bilateral (N = 77) mTESE from 2018 through 2019. One or more biopsy samples representative of the overall appearance of the testicular parenchyma were taken from one (for unilateral mTESE) or both testes (for bilateral mTESE), was fixed in Bouin’s solution and sent to the pathologist. Histological analysis was conducted by the same experienced pathologist, who examined at least 100 different tubule sections per biopsy sample. Participants/materials, setting, methods Each tubule section (N = 100 per sample) was evaluated at 10X magnification with a micrometer to measure the tubule diameter, then the mean ST diameter was computed. The basal membrane (BM) thickening was evaluated in every section, and a score was assigned by multiplying the degree of BM fibrosis (mild=1, moderate=2, severe=3) for the number of sections (e.g. BM score for moderate fibrosis in 50 sections=2x50=100). Leydig cells hyperplasia (LCH), if present, was also annotated. Main results and the role of chance The median + interquartile range STs diameter was 140; 110–185 mm, while the median BM score was 100; 10–150. Sertoli cell only syndrome (SCO) was found in 51.1% of cases, focal SCO (FSCO) in 4,7%, early (EMA) and late (LMA) maturation arrest in 10.2 and 2.73% respectively, HYPO in 26.17% and hyalinosis (HL) in 5% of cases. LCH was found in 46.88% of samples. STs diameter, BM score and LCH differed significantly among the different histological patterns: STs diameter was 125; 100–148 in SCO, 162; 102–187 in FSCO, 130; 100–175 in EMA, 145; 130–195 in LMA, 205; 170–240 in HYPO and 57.5; 42.5–100 in HL. HYPO samples also had the lowest BM score (20;1–100; p < 0.0001) and LCH prevalence (23.8%, p < 0.0001) compared to the other histological subgroups (HS). A multinomial logistic regression for prediction of different histological subgroups was run with STs diameter, BM score and LCH as candidate predictors: the model explained the 29% of variability of the outcome and correctly classified 69% of cases. STs diameter significantly predicted FSCO (RR 1.02, 95% CI 1.0–1.04), LMA (RR 1.02, 95% CI1–1.04) and HYPO (1.03, 95%CI 1.02–1.04), while BM score significantly predicted HL (RR 1.07, 95%CI 1.02–1.13). Limitations, reasons for caution The STs were carefully cut before extraction in order to preserve their structural integrity, however the accuracy of such a method of estimating the STs diameter needs to be assessed by further studies. Wider implications of the findings: The identification of dilated STs remains the best strategy to retrieve sperm by mTESE, since larger STs diameters are mostly associated with the more favorable histological patterns. However, dilated STs may be also found in cases with LMA, which explains why in some cases dilated STs do not contain sper. Trial registration number Not applicable