Conformal radiation therapy (CRT) and intensity-modulated radiation therapy (IMRT) are used to precisely deliver radiation to the prostate while sparing surrounding organs. However, both methods are limited by the reality that the prostate's position, size, and shape change from day to day -- and even within a single treatment session -- due to variations in bladder and rectal filling.
The traditional solution is to add a large planning target margin of 10 mm or more around the prostate, so that even if the organ shifts, the cancer still receives adequate dose. But this margin expands the treatment volume to include nearby healthy tissue, increasing radiation dose to the rectum and bladder and raising the risk of bowel and urinary side effects.
Two clinical studies demonstrated the real consequences of positional uncertainty. De Crevoisier et al. found that patients whose planning CT showed a distended rectum experienced a nearly 30% decrease in biochemical control at 5 years, because during treatment the rectum was less distended, shifting the prostate posteriorly and causing systematic target underdosing. A Dutch study confirmed a 20% reduction in freedom from failure in patients with larger rectal volumes.
Adaptive radiotherapy was developed to address this fundamental limitation by personalizing each patient's treatment plan based on actual measurements of their individual anatomical variation captured during the treatment course, rather than relying on generic population-based margins.
The off-line adaptive process developed at William Beaumont Hospital begins with a standard treatment plan for the first 5 fractions. During these initial fractions, patients receive daily portal imaging and serial CT scans on each of the first 4 treatment days. These measurements capture both systematic setup errors (consistent positional bias) and random day-to-day variation in organ position.
After analyzing these measurements, a patient-specific planning target volume (PTV) is constructed. This custom target volume accounts for the individual patient's measured systematic error -- essentially correcting for a predictable offset -- while still including margin for the random variation characteristic of that particular patient. This replaces the generic population-based margin with one tailored to the individual.
Using this personalized target, a second modified treatment plan is calculated and implemented for the remaining fractions. The prescription dose is selected individually for each patient based on how much radiation the rectum and bladder can safely receive, as determined by dose-volume histogram constraints. Patients whose anatomy allows higher doses receive them; patients at higher risk of organ toxicity receive appropriately reduced doses.
A key advantage of this approach is that the dose to the prostate can often be escalated safely. Because the target margin is reduced, the beam can be tightened around the prostate, decreasing the dose to the rectum and bladder -- which in turn allows the total dose to the prostate to be increased without exceeding toxicity thresholds. Studies showed average dose escalation of 3.6 Gy with CRT and 5.4 Gy with IMRT compared to conventional treatment.
Long-term data from 3,064 prostate cancer patients treated at William Beaumont Hospital demonstrates a dramatic improvement in biochemical control -- measured by PSA levels after treatment -- over successive treatment eras. The 5-year biochemical control rate was 51% for patients treated in 1987-1990, rising to 62% in 1991-1995, 83% in 1996-2000, and 90% in 2001-2005, corresponding to the era of dose-escalated adaptive radiotherapy.
Similarly, the 5-year clinical failure rate fell from 24% in 1987-1990 to 2% in 2001-2005 (p less than 0.01). These improvements correlated with the introduction of image-guided adaptive treatment and higher prescription doses, alongside improvements in overall and cause-specific survival. While multiple factors contributed, adaptive radiotherapy enabling safe dose escalation was a central driver.
A dedicated toxicity analysis of 642 patients treated with the adaptive process and followed for at least 4 years showed that grade 3 toxicities were infrequent, generally below 5% across all dose levels ranging from 70.2 to 84 Gy to the isocenter. Crucially, there were no statistically significant differences in chronic toxicity rates among the three dose levels studied (low, intermediate, and high dose), confirming that higher doses could be delivered safely when guided by adaptive planning.
A comparison of adaptive CRT versus adaptive IMRT in 728 patients showed IMRT substantially reduced side effects. Chronic rectal bleeding at grade 2 or higher occurred in 16% of adaptive CRT patients but only 4% of adaptive IMRT patients (p less than 0.01), with the IMRT group also treating more high-risk patients with larger target volumes, making the toxicity advantage even more striking.
While off-line adaptive radiotherapy modifies the plan between treatment sessions, online image-guided adaptive radiotherapy uses imaging performed immediately before (or during) each fraction to adjust the treatment in real time. The key enabling technology is cone-beam CT (CBCT), which is built into the radiation delivery machine and can scan the patient while they lie in the treatment position.
Online adaptation strategies range from simple to sophisticated. The simplest approach adjusts patient position to align the prostate with the planned location. More advanced methods modify the beam aperture shape or multi-leaf collimator (MLC) segment positions to match the prostate's actual shape and position that day. The most sophisticated approach is full online inverse planning, which reoptimizes the entire treatment plan based on the daily image.
Online inverse planning studies estimated an average additional target dose increase of 13% compared to position correction alone, based on equivalent rectal dose. However, responses varied widely: 27% of patients had minimal benefit (less than 5% dose increment) and 32% had large benefit (more than 15% dose increment), emphasizing that personalization matters and some patients benefit far more than others.
A critical challenge for online adaptive techniques is extra-capsular extension (ECE) -- microscopic cancer spread beyond the visible prostate capsule. Analysis of surgical specimens found that more than 4 mm of ECE could be present in 20% of patients with PSA above 10 ng/mL and high Gleason scores. Reducing treatment margins to 1-3 mm in such patients risks missing this occult disease, requiring careful patient selection before extremely tight margin approaches are used.
Even within a single treatment session, the prostate can move as the patient lies still -- due to breathing, bowel peristalsis, and muscle relaxation. Studies of 22 patients found that 77% had intra-fraction motion requiring only a 3 mm margin for coverage, while 23% had larger motion requiring 6 mm margins, highlighting significant patient-to-patient variability in movement during treatment.
Seminal vesicles -- which are often included in the treatment target for high-risk patients -- move more than the prostate itself, requiring approximately 50% extra margin for the seminal vesicles compared to the prostate when targeting both structures during online image guidance. This complicates margin reduction strategies for patients who need seminal vesicle coverage.
A hybrid strategy has been proposed that combines the strengths of both approaches: online position correction is performed daily, while patient-specific intra-fraction motion margins are measured during the first 4 fractions and incorporated into a modified planning target volume for the remaining treatments. This accounts for both day-to-day variation and within-session motion specific to each patient.
Adaptive radiotherapy represents a fundamental shift from population-based to personalized radiation treatment. By measuring each patient's actual anatomical variation and incorporating it into the treatment plan, margins can be safely reduced -- allowing dose escalation that improves tumor control while limiting damage to the rectum and bladder.
Current onboard CBCT imaging has limited soft-tissue contrast, making it difficult to precisely identify prostate boundaries during online guidance. Onboard MRI machines integrated with radiation delivery systems (MR-Linac technology) are emerging as a transformative solution, offering real-time high-quality soft tissue visualization that could enable more accurate and responsive online adaptation.
The authors emphasize that new technology should be implemented cautiously and comprehensively. Key uncertainties that must be addressed include extracapsular extension -- the presence of cancer beyond the visible prostate edge -- and the need to validate that reduced margins do not compromise cancer control for patients with more extensive disease.
With approximately 2,000 patients treated successfully using off-line adaptive prostate radiotherapy at the time of this review, and with continuous improvements in imaging, planning algorithms, and delivery technology, adaptive radiotherapy is positioned to become a clinical standard that simultaneously maximizes cancer cure and minimizes treatment-related harm.