Comparison of Radiation Side-Effects of Conformal and Conventional Radiotherapy in Prostate Cancer: a Randomised Trial

Dearnaley DP, ... Horwich A.
Abramson Cancer Center of the University of Pennsylvania
Last Modified: November 1, 2001

Reviewers: John Han-Chih Chang, MD
Source: Lancet 1999; Volume 353: pages 267 - 272


It has been known that radiation could eradicate tumors -- the radiation treatments just needed to utilize a "high enough" dose. Theproblem with tumoricidal doses of external beam radiation is that normal tissues can also be affected. This leads to acute andlong-term toxicities of treatment. In today's world of advancing technology, three-dimensional conformal radiation therapy (3D-CRT)planning systems are able to generate treatment schemes that are supposedly conformal to the tumor/target volume thereby limitingthe dose that reaches the surrounding normal tissues. This would, thus, allow the radiation oncologists to increase the doses to thetarget (the prostate, in this case), while minimizing dose to the normal tissues in proximity (bladder, intestine and rectum). In theory,this is a wonderful concept. It has been somewhat confirmed in nonrandomized trials of 3D-CRT compared with historical controls.

There have been very little in the way of randomized trials of conventional versus 3D-CRT. The British attempted to venture downthis road less traveled. A randomized study was initiated at the Royal Marsen NHS Trust and Institute of Cancer Research in theUnited Kingdom and is the subject of this article.

Patients and Methods

Their initial trial was for pelvic cancers comparing 3D-CRT versus conventional RT. Of 274 patients, 144 were prostate cancerpatients and were included in this analysis with an additional 98 added when the trial included prostate cancer patients alone. Thus,242 were eligible for treatment. Seventeen were excluded, because they were treated for palliation only. Of the remaining 225, 111were randomized to receive conventional RT, while the rest got conformal therapy.

Clinical stages included in this trial were any primary tumor extent without lymph node or distant metastases (although one patientin the conventional treatment arm did have regional lymph node metastases). Table 1 demonstrates the patient characteristics.Though minimally different, The conformal therapy arm did have more T1 and less T2 and T4 lesions than the conventional RT arm.There were also lower grade lesions in the conformal treatment arm in contrast to the conventional RT arm. The median PSA washigher in the conventionally treated patients.

The field arrangements for the RT portals utilized a 3-field technique with an anterior and laterals or posterior obliques. Computedtomography (CT) scans were utilized to localize the prostate gland. Customized cerrobend blocks were created to ensure that the90% isodose line covered the prostate gland in the 3D-CRT arm (1.0 cm margin around the prostate and seminal vesicles defined thePTV and a 0.6 cm margin was used from the PTV to the block edges). Treatment included the base of the seminal vesicles in mostcases, but poorly differentiated cancers and locally advanced lesion also received treatment to the entirety of the seminal vesicles.The conventionally treated arm did not utilize any blocks. All patients were treated to 60 - 64Gy with 2Gy fractions per day. Six toten megavoltage photons delivered the treatments, which were prescribed to isocenter. In over half of the patients, androgenablation was administered 3 - 6 months prior to RT.


Complete 2-year data was obtained on 99 3D-CRT patients, while 94 conventional RT patients had full 2-year follow-up. The rate ofsevere (at least grade 3 -- symptoms affecting lifestyle and may require hospital admission or minor surgical intervention) was verylow in both arms and appeared not to be statistically significant. Grade 2 proctitis (rectal pain and/or bleeding that is responsive tooutpatient medical treatment and does not affect lifestyle) was more often seen in conventional RT patients. The actuarialprobability of remaining free of grade 2 or higher proctitis at 5 years was 92% compared to 82% for conventional RT. No significantdifference in other late gastrointestinal (diarrhea) or genitourinary (hematuria, cystitis or incontinence) effects was seen (detailed inTable 2).

Overall survival was similar for the two groups at approximately 90% at 2 years and 65% at 5 years. The clinical local control rate wasapproximately 97% versus 96% at 2 years and 78% versus 83% at 5 years for 3D-CRT versus conventional RT, respectively. Thebiochemical disease control rates were based on a cutoff system -- if the PSA rose above 2 ng/mL, it would be considered a failure.There seemed to be an advantage to receiving 3D-CRT for biochemical disease control based on that definition of biochemicalfailure. However, since the median PSA is higher in the conventional group, it stands to reason that the conventional RT group,being that they started with a higher baseline, would be above the cutoff more often than the 3D-CRT group. As described by theauthors, when they stratified for initial PSA there was no significant difference in biochemical local control.

Discussion and Critique

Many have touted this article as first randomized trial to prove that there was an advantage to 3D-CRT in the treatment of localizedprostate cancer in minimizing long term toxicities. However, there are major flaws that confound the issue. The first and foremostproblem is the fact that this is not a trial of conventional versus 3D-CRT, but rather of blocked (standard of care utilizing 1.5 - 2.0 cmof margin on the target volume) versus unblocked fields (which is considered substandard care in this day and age). That questionhas been intuitively obvious (blocked is better -- now we have a randomized trial stating that). Another argument was thestratification of the randomized patients. As mentioned above, the conventional RT patients had a higher median PSA, grade, andclinical stage of tumor than the 3D-CRT arm. This would lead to patients receiving a larger volume of treatment, compromising theamount of bladder and especially the intestines and rectum that could be spared in the unblocked fields of "conventional" RT.

Additionally, average field dimensions were not given for the reader to ascertain how truly different the fields were. Anotherproblem I have with the paper is the fact the dose of RT was substandard to the periphery of the gland. If 64Gy was prescribed toisocenter and the 90% isodose line covers the prostate, then the periphery of the gland only received 57.6Gy (substandard). Theauthors discuss dose escalating to 74Gy as the focus of the current trial --this would deliver 66.6Gy to the periphery (nearly up tothe standard dose used for localized prostate). The issue of using grade 2 toxicity as the criteria for the advantage of more"conformal " RT is a rather weak argument, since most agree grade 3-5 toxicities are the dose-limiting ones (not significantly differentfor any of the side effects of treatment in this trial). Grade 2 toxicities, while they are significant, are usually reversible with outpatienttreatments and are not dose limiting.

My final contention is that the biochemical failure criteria utilized was a problematic approach to this trial since there was adiscrepancy in the initial PSA between the two arms. In conclusion, this article has demonstrated that we must be critical of theliterature, despite of what may be the prevailing message that the authors are attempting to convey. I am not disparaging conformaltherapy. I feel we do not have all the answers yet. 3D-CRT is probably no worse than conventional RT and most likely isadvantageous, but a randomized trial has yet to definitively prove that fact.