Triple-negative breast cancer (TNBC) occurs in about one-sixth of all BC patients, with the most aggressive behavior and the worst prognosis. It is a heterogeneous disease, with specific molecular characteristics and natural dynamics of early relapse and rapid progression. In the absence of a valid biomarker, and thus any targeted treatment, chemotherapy is nowadays still the backbone of treatment, with the so-called. TNBC paradox - better chemosensitivity and response to therapy, but worse survival. Today, more hope lies in finding new biomarkers that will be predictive and prognostic in TNBC, such as androgen receptors (AR) and tumor-infiltrating lymphocytes (TIL). Retrospective cohort study was conducted on a consecutive sample of 152 patients diagnosed and treated for early TNBC at the University Hospital for Tumors, Sestre milosrdnice University Hospital Center, from Jan 1st 2009 untill Dec 31st 2012. Sociodemographic and clinicopathological variables available from the hospital record were analyzed and IHC AR analysis and TIL morphological analysis by HE were performed. Limit for AR positivity was ≥1%, and TILs were fully analyzed, stromal (sTIL) and intratumoral (iTIL), in the central tumor (CT) and at the invasive margin (IM). Correlations of all parameters were calculated as well as 5y DFS and OS. Patients were at median age of 58, mostly postmenopausal, most commonly stage II disease, grade III, and a median Ki-67 of 57%. They were treated surgically and then in almost 90% of cases with adjuvant chemotherapy, mainly anthracyclines and taxanes and adjuvant radiotherapy. At 5y follow-up, disease recurrences are reported in one third of patients, most often in the first year after diagnosis, mainly in the form of distant metastases to the lungs and bones. The 5y DFS was 67.1% and the OS was 73.7%. Factors such as younger age, larger tumors, involved lymph nodes, or a higher stage of disease, and a higher Ki 67 were most significantly associated with disease recurrence and death. AR were expressed (≥1%) in 31.1% of patients, and in 18.5% of them AR were> 50%. By evaluating TIL according to the above mentioned technique, sTIL on IM were most frequently represented, with a median intensity of 30%, and as many as 85.5% of patients with ≥10%, while the rarest and of lowest intensity were iTIL in CT, with a median intensity of 1% and 23% of patients with ≥10%. In total, quarter of patients had TIL>50% in any of the evaluated compartments. Around 30% of patients had expressed AR (≥1%) and present TIL(>1%) at the same time, and AR expression was not statistically significantly associated with the presence of TIL. Expression of AR showed no statistically significant effect on DFS, or OS. Evaluation of TIL showed a statistically significant favorable effect on DFS of sTIL on IM, but not of the other compartments. Overall, patients with TIL≥10% in all four evaluated compartments (sTIL and iTIL, in CT and IM) were shown to have statistically significantly better DFS, compared to those with TIL<10%. Stromal TIL had statistically significant effect on OS, iTIL only in the case of ≥10% intensity, sTIL on IM were statistically significant predictor of better OS, and iTIL on IM only in the case of ≥10% intensity. AR and TIL in the survival analysis showed that the presence of iTIL in CT was significantly negatively associated with 5y OS in AR+ tumors, whereas presence of TIL≥10% and AR≥1% was not statistically significant, independent predictor of OS. AR were found to be statistically significantly positively correlated with age and menopausal status, and negatively with tumor size, grade and Ki-67, as well as disease stage. Stromal TIL in CT were statistically significantly associated with age and consequently menopausal status and tumor size, but without correlation with other variables, such as histologic subtype, lymph node status, gradus, and Ki-67; iTIL in CT were statistically significantly positively associated with histologic subtype, ie NOS, but also negatively with age and menopausal status, exactly opposite to sTIL in this section, and also without association with other tumor characteristics, such as tumor size, status lymph nodes, gradients or Ki 67; sTIL on IM, as well as iTIL on IM, showed statistically significant association with tumor grade and Ki-67, and no association with age and menopausal status. In this analysis, AR did not prove to be an independent factor in the prognosis of patients with early TNBC. TIL, as described above, proved to be a statistically significant independent predictor of prognosis, with the result expected to be mostly at the expense of stromal TIL. Evaluation of TIL by individual compartments indicated the prognostic value of sTIL on IM. This analysis also showed that the potential additional prognostic value lies in the interaction of AR with sTIL and iTIL on IM, although the results are of very low precision and reliability. Our survival outcomes for patients with early TNBC are consistent with similar analyzes in cohorts of patients from neighboring as well as distant geographic areas. As in most analyzes, the prognostic impact of AR did not prove statistically significant. For the first time, TIL were analyzed with sTIL and iTIL, in CT and on IM, and a probably different, prognostically significant role of a particular morphological compartment in the tumor, especially IM, was observed. By far the most interesting part of the analysis was the correlation of otherwise independent parameters, AR and TIL, of which little is known today from sporadic reviews on the effect of immunity on molecular subtypes of TNBC, and which still lacks a definitive conclusion, which is why also this analysis, providing interesting preliminary results, requires validation in a larger, prospective study.