engleski

Artificial intelligence, active surveillance and treatment of localized prostate cancer

Treatment landscape of localized prostate cancer is rapidly changing. Both radical prostatectomy and radiotherapy as two main treatment modalities in this setting have adopted new technologies that optimally would translate into more precise treatment with better cure rates and less side-effects. Robotassisted radical prostatectomy, patient-friendly alternative to open retropubic prostatectomy, is gaining momentum worldwide for prostate and lymph nodes removal and has become standard-of-care. Radiotherapy is nowadays coupled with unprecedented imaging and computing capabilities allowing millimeter- level accuracy and safe delivery of larger doses per fraction, enabling shorter courses of previously weeks-long treatment schedules for localized prostate cancer. There is still ongoing competition between radiotherapy and prostatectomy in localized prostate cancer. As previously discussed, both treatments changed and improved over the last decade. To complicate things further, active surveillance emerged as increasingly utilized third option in localized disease management. In this context the only randomized trial that assessed long-term efficacy and side-effects profile of these treatment options for localized prostate cancer is The Prostate Testing for Cancer and Treatment (ProtecT) randomized clinical trial. ProtecT trial recruited 1643 men aged 50–69 years with clinically localized prostate cancer and randomized them to active monitoring, radical prostatectomy and radical radiotherapy. The primary intention-to-treat analysis at a median follow-up of 10 years showed that the rate of overall mortality was ~1%, irrespective of treatment assigned. However, radical treatment was associated with ~50% reduced disease progression/ metastasis compared with active monitoring (active monitoring 6.3, radiotherapy 3.0 and radical prostatectomy 2.4 events per 1000 person‐years ; p = 0.004). With the progression of prostate cancer radiotherapy the idea was born that hypofractionation can bring clinical benefit based on the assumptions of low alpha/beta ratio for prostate cancer, which was con7 Lib Oncol. 2023 ; 51(Suppl 1):1–44 sidered to be in the range of 1.5 Gy. However, mature trial results widely refuted theoretically anticipated hypofractionation advantage (i.e. superiority). Actually, what was largely observed at best is the noninferiority of hypofractionation compared to standard fractionation. Furthermore, analysis showed that there might be dose-threshold effect on dose-response curve. With radiotherapy equivalent doses (EQD2) going beyond 80 Gy it seems the gain associated with further dose escalation would be marginal at best, suggesting plateau of the dose-response curve for doses ≥80 Gy While there is limit of safe whole prostate dose escalation with external beam radiotherapy, there is potentially a therapeutic gain if we specifically target and boost portion of the prostate containing the bulk of the tumor, are called dominant intraprostatic lesion (DIL). Wide use of MRI allowed to dissect prostate anatomy and to clearly visualize intraprostatic tumors which enabled their targeting with radiotherapy. Authors performed stereotactic boost to DIL using volumetric modulated arc therapy and reported early efficacy and toxicity endpoints in the recent JCO publication. In the FLAME trial the authors hypothesized that focal boosting of the macroscopic visible tumor with external beam radiotherapy would increase biochemical control in patients with localized prostate cancer. They recruited 571 patients with intermediate- and high-risk prostate cancer and randomized them to either standard treatment (77 Gy in 2.2 Gy daily fractions) or to the addition of integrated simultaneous tumor boosting to 95 Gy (2.7 Gy fractions, EQD2=115.8 Gy). Majority of patients received long-term ADT. Reported 5-year biochemical control was 92% in focal boost arm and 85% in the standard arm (p<0.001) without additional toxicity associated with focal dose escalation. Altogether, horizons of radiation therapy in localized disease are changing adopting multimodal treatment such as focal mpMRI guided radiotherapy boosts, external beam radiotherapy with brachytherapy boosts, extreme hypofractionation for appropriately selected patients, consideration for pelvic lymph node irradiation in men at high risk for nodal involvement. Radiotherapy is becoming more precise and personalized. Active surveillance (AS) is nowadays viable option for patients with low-risk localized prostate cancer. It emerged because of the clear and obvious fact that today more patients are diagnosed with prostate cancer than ever before and many of them are over-diagnosed and overtreated. More than 6 times of the patients are diagnosed with prostate cancer, then the number of patients who dies from it at the end. AS consists of periodical MRI scans, prostate biopsies and biochemical surveillance (PSA blood test) to catch ‘’upgrade’’ of the prostate cancer in due time. The pros for this kind of watchful waiting is evading potential short and long-term side effects that can emerge after radical prostatectomy or during and after prostate cancer radiotherapy and prevention of overtreatment in patients who will, maybe, never need active treatment. The cons for this kind of prostate cancer management are that there is still no consensus on optimal AS protocol and inclusion criteria for appropriate low-risk patients. Unfortunately, today there is still no effective predictive biomarker who can select appropriate patients with local prostate cancer that can benefit from specific kind of available treatment options. This is especially seeable when trying to decide which patients with intermediate or high risk localized prostate cancer will benefit from radiotherapy alone or radiotherapy + ADT combined. Dan Spratt published very interesting results where his team trained and validate first ever predictive biomarker that can help assess patients with prostate cancer who can clearly benefit from ADT use. They used pretreatment biopsy slides from multiple phase III NRG Oncology randomized trials of men receiving RT or RT+ADT. At the end they have managed to train artificial intelligence (AI) to derive predictive biomarker from screening pretreatment biopsy slides and was trained to predict distant metastases. The results showed that 15-year distant metastases rate difference between RT versus RT+ADT in the biomarker negative group was 0.3%, vs biomarker positive group 9.4%. This clearly shows the potential benefit of using AI in patient treatment deciLibsion making, which can lead to prevention of overtreatment and lower the overall cost of potentially not needed therapy.

prostate cancer, personalized radiotherapy, focal boost, active surveillance, artificial intelligence, androgen-deprivation therapy, hypofractionation

nije evidentirano