In this study two different 3DCRT techniques (WP and PO) were analyzed and the degree of association was determined between the occurrence and evolution of acute and late GI and GU toxicities and the treatment related characteristics in patients entering our hospital. Important findings include: (i) a higher proportion of acute GI toxicity in the WP 3DCRT technique group and conversely a slightly higher proportion of late GI and GU toxicity in the PO patient group; (ii) acute GI toxicity as a significant predictor of late GI toxicity; (iii) a strong dependence of the occurrence and evolution of acute GI toxicity and of late GU toxicity on which 3DCRT technique is used; (iv) the association of both acute and late GU toxicity and radical prostatectomy performed prior to radiotherapy; (v) the influence of age on both acute GI and GU toxicities; (vi) a correlation between the percentage of volume of irradiated bone marrow and a decreased number of leukocytes; and (vii) the influence of radiotherapy preferentially on NK, NKT and T cell subpopulations.
We found an increase of acute vs. pretreatment GI symptoms predominantly in the WP group, even if the patients were irradiated with lower doses compared with the PO 3DCRT group. We assume that the limiting factor in high-volume irradiation is not the dosimetric parameters, but the overall patient tolerance. In addition, the WP technique was undergone by patients with advanced stages of disease, lower overall health status, and suppressed immune functions. These observations are supported by data of Jereczek-Fossa  and Schultheiss et al. ; however, some investigators didn't demonstrate this correlation . On the other hand, the diminution of late GI and GU toxicities to grade 1 or to no toxicity in the majority of acute toxicity (grade 1-3) suffering patients, was observed also in the WP 3DCRT group.
Our data regarding the frequency of severe toxicities are similar to those of other series, despite the fact that a direct comparison of toxicities is difficult due to the existence of many modified versions of the classification, and modifications of grading scales. Similarities were found between our results, the RTOG 9413  analysis, and the GETUG-01  prospective study. The diversity in the diagnostics could be created by individual physicians due to the subjectivity of the scoring system, when the same toxicity could be graded differently. Due to the findings of decreased late GI and GU toxicities after 3DCRT in the cohort of our patients, we compared these results with the studies using hypofractionated stereotactic body radiotherapy SBRT, which is a new modality of localized prostate cancer RT. The SBRT, together with innovations in image-guidance technology, is able to automatically correct the movement of the prostate during treatment, and deliver highly-conformal beam profiles, which have greatly enhanced the capability of delivering high dose fractions to a well-defined target, with sharp dose fall-off towards the bladder and rectum. Most of the studies concerning SBRT as a monotherapy or even as a boost following external beam radiotherapy presented only negligible incidence of severe late GI and GU toxicity. Katz et al. [24, 25], Freeman et al.  and other authors  reported milder toxicity profiles in comparison with our results, particularly in the case of late GI and GU impairments. On the contrary, Jabbari et al.  presented in their study similar results as our ones in the occurrence of severe late GU toxicity (grades 2-3), and even worse outcomes in evolution of acute GU toxicity.
The analysis of GU toxicity is difficult, due to interference with pre-existing dysfunction, age-related diseases, and previous urological surgery [21, 29]. We have to remember that some of these pre-existing symptoms could have been erroneously interpreted as acute or even late GU toxicity. On the other hand, late bladder damage can occur with a long latency time, potentially resulting in the underestimation of the real severity of late toxicity . The difference in the time of clinical manifestation could be the reason why some researchers demonstrated the correlation of acute and late GI, but not GU toxicity . These outcomes were proved in our study, as well.
The development of acute 3DCRT- induced GI and/or GU damage was generally mild in both groups; and none of the patients had an interruption of radiotherapy due to toxicity side effects. The risk of both acute GI and GU reactions depended preferentially on the age of patients, in agreement with the results demonstrated by Jereczek-Fossa et al. . The biological variables and different clinical decisions based on patient age could participate on the final outcome. The association of acute GU toxicity with the percentage of the urinary bladder receiving 50 Gy found in our study was in accordance with the results of Fiorino et al. and other authors [21, 29], who reported a significant correlation between DVH parameters and incontinence. Similarly, the acute GI toxicity associates with the percentage of rectum receiving 70 Gy as demonstrated also by the Italian Association for Radiation Oncology (AIRO) Group on Prostate Cancer (AIROPROS) 0101 trial (previous retrospective investigation , who described that the dose of 70 Gy at rectum was predictive for late G2-G3 bleeding), AIROPROS 0102 , and by others [[34, 30, 7, 35]]. Both late GI and GU toxicities positively associated with the volume of the irradiated organ at risk, rectum and urinary bladder, respectively. Furthermore, the late GI toxicity is associated with stage T of TNM classification of the disease, and is strongly influenced by acute GI toxicity. These finding are in agreement with the published data of Heemsbergen et al. . The risk of late GU reactions depended on the percentage of urinary bladder volume receiving 40 Gy, the 3DCRT technique used, and the previous urological surgery [21, 29].
Originally, the primary mechanism of RT in cancer reduction has been considered the neoplastic cell DNA damage. However, Takeshima et al. have found that tumor-specific CTL, which were induced in the draining lymph nodes and tumor tissue of mice by RT, are fundamental to the inhibition of cancer growth . The immunological evaluation performed during 3DCRT showed a positive correlation of the number of activated NK cells and the proportion of terminally differentiated tumor targeted cytotoxic effectors with GI and GU toxicities. Both of these subpopulations returned to normal values or decreased after completing RT. In contrast, T lymphocytes were decreased during RT and normalized after its completion; while NKT cells were down-regulated in all time periods. The acute GU and late GI and GU toxicities significantly increased the T cell proportion, NK cell-mediated cytotoxicity, and cytotoxic T cell numbers. We assume that these changes are caused by stress conditions induced by RT-damaged and GI or GU toxicity-affected tissues, eliciting stimulation of cytotoxic cells (NK and CTLs). These RT effects could be due to inflammation following increased apoptotic/necrotic events in the involved tissues. The surface expression or extracellular release of stress proteins (e.g. MICs, Hsp70), following tumor cell damage by RT, can play a key role in immune system modulation . These molecules are ligands of the NK cell activation receptor NKG2D , and can stimulate NK cell functional maturation. Particularly, Hsp72 can act as an immunological adjuvant [39, 40], participating in the non-self recognition of prostate cancer cells. Thus we can hypothesize, according to results of Hurwitz et al.  that the enhanced immune function, involving resting and terminally differentiated NK cells during 3DCRT, as well as the up-regulation of CTL number and the NK cell-mediated cytotoxicity in GI or GU suffering patients, could follow the release of HSPs either evoked by radiation or by GI or GU toxicity-induced cellular stress.