Improvement in toxicity in high risk prostate cancer patients treated with image-guided intensity-modulated radiotherapy compared to 3D conformal radiotherapy without daily image guidance

Radiotherapy in combination with androgen-deprivation therapy (ADT) is a well-documented treatment option for high risk prostate cancer (PCa) [1]. With intensity modulated radiotherapy (IMRT) it is possible to deliver high doses to the target and at the same time reduce the dose to the surrounding normal tissue [2]. Recently, the concept of image guided radiotherapy (IGRT) has been introduced in the treatment of PCa. IGRT is often based on the implantation of fiducial gold markers in the prostate. With markers implanted in the prostate, the position of the prostate can be verified before each treatment fraction using portal imaging. This limits the interfractional variability in the position of the prostate and as a consequence the PTV margins can be reduced [3]. In theory, a reduction in PTV margins should reduce the dose to the organs at risk and consequently result in less toxicity [4]. However, there is concern that large margin reductions might increase the risk of geographical miss of the target, which might result in higher rates of local failure [5].

In our centre, the RT for PCa has changed during the past decade. In 2000–2005 patients were treated with 3D conformal radiotherapy (3DCRT) to 76 Gy and without daily image guidance. In 2005, we introduced daily image-guided intensity modulated radiotherapy (IG-IMRT) based on implanted markers in the prostate. With the introduction of IG-IMRT, the PTV margins were reduced substantially, and the dose was increased to 78 Gy.

The purpose of this study is to compare toxicity and biochemical progression-free survival between patients treated with IG-IMRT and 3DCRT for high risk PCa. We hypothesize that IG-IMRT and reduced margins result in less toxicity compared to 3DCRT with broader margins and without daily image guidance. On the other hand, given the accuracy of IGRT the reduction in margins should not compromise the level of tumor control in IG-IMRT patients.

The 2-year actuarial likelihood of developing grade ≥2 GI toxicity following RT was 57.3% for 3DCRT and 5.8% for IG-IMRT (p < 0.001). For grade ≥2 GU toxicity the corresponding numbers were 41.8% and 29.7% (p = 0.011), respectively (Figure 1).

Predictors of toxicity and biochemical progression are shown in Tables 2 and 3, respectively. For the continuous variables age, PPB, Gleason score and PSA, the results indicate the hazard ratio per one unit increase in the variables. For PS the results indicate the hazard ratio for patients with PS = 1 compared to patients with PS = 0. For the T-stage the results indicate the hazard ratio for patients with T-stage ≥ T2c compared to patients with T-stage ≤ T2b.

In the multivariate analysis, 3DCRT was associated with a significantly increased risk of developing grade ≥2 GI toxicity compared to the IG-IMRT group (p < 0.001, HR = 11.59 [CI: 6.67-20.14]). 3DCRT was also associated with a significantly increased risk of developing grade ≥2 GU toxicity compared to IG-IMRT (p = 0.015, HR = 1.72 [CI:1.11-2.67]). Furthermore, patients with comorbidity had a higher risk of developing GU toxicity compared to patients without comorbidity (p = 0.038, HR = 1.50 [CI:1.02-2.21]). Finally, the use of ADT was protective of GI toxicity (p < 0.001, HR = 0.23 [CI: 0.10-0.50]).

The 3-year actuarial biochemical progression-free survival probability was 86.0% for 3DCRT and 90.3% for IG-IMRT (p = 0.386) (Figure 2). As shown in the Cox regression analysis in Table 3, there was no significant difference between 3DCRT and IG-IMRT in the risk of developing biochemical progression (p = 0.505, HR = 0.78 [CI:0.37-1.57]). Increasing age was associated with a lower risk of developing biochemical progression (p = 0.017, HR = 0.93 [CI: 0.88-0.99]). Higher percentage of positive (malignant) biopsies (PPB) was predictive of biochemical progression (p = 0.038, HR = 1.01 [CI: 1.00-1.02]).

In this study, we compared toxicity and biochemical progression-free survival in high risk PCa patients treated with either of two RT techniques: 3DCRT without daily image guidance versus daily fiducial-based IG-IMRT. IG-IMRT was introduced in our centre in 2005 and was accompanied by a small dose escalation and a substantial reduction in PTV margins. To our knowledge, this is one of the largest studies to demonstrate the clinical effect of introducing IG-IMRT and smaller margins in the treatment of high risk PCa. In lack of randomized trials comparing RT techniques and different margins, it is important that we keep reporting clinical outcomes following the implementation of new radiation techniques [12].

We found that 3DCRT patients experienced significantly more toxicity than IG-IMRT patients. The observed difference in GI toxicity was large and can be attributed to the combination of the accuracy of IGRT, reduced PTV margins and IMRT-based dose sparing of the rectum [4, 13]. Furthermore, the use of MR imaging in the treatment planning of IG-IMRT probably results in reduced dose delivery to the organs at risk due to smaller clinical target volumes [14]. We speculate that the major contributing factor for the large difference in GI toxicity is the large difference in PTV margins between the two groups. An example of a 3DCRT plan with 2 cm PTV margin and a fixed-angle IMRT plan with 5–7 mm PTV margins is shown in Figure 3. Assessment of the differences in dose-volume parameters between the treatment groups was beyond the scope of this clinical report.

Compared to other studies of non-image guided 3DCRT, we observed a relatively high degree of GI toxicity in our 3DCRT patients. However, our results are comparable to other reports of 3DCRT using the BeamCath technique [9, 15]. Furthermore, the 2 cm PTV margin used in most fractions in 3DCRT is wider than what is commonly used in studies of 3DCRT without daily image guidance. The difference in GU toxicity between the two treatment groups was not nearly as significant as the difference in GI toxicity. A part of the explanation might be that the prostatic part of the urethra was not spared in either of the treatment groups [16]. Not surprisingly, patients with comorbidity seemed to have a higher risk of developing GU toxicity compared to patients without comorbidity.

Even though the entire cohort consisted of high risk patients, 30 patients did not receive adjuvant ADT. Interestingly, we found that the use of adjuvant ADT was independently associated with a lower risk of developing GI toxicity. Although short-term ADT has previously been shown to be protective of toxicity [17], the actual mechanism of such an effect remains unclear.

There was no significant difference in biochemical progression-free survival between the two treatment groups. The minor difference in prescribed dose was not expected to influence the biochemical progression-free survival to a large degree [18]. The major difference between the two groups besides the actual radiation technique was the use of daily image guidance and the size of the PTV margins. The 2 cm margin used in the majority of fractions with 3DCRT should be large enough to account for the interfractional and intrafractional motion of the prostate [19]. With the introduction of IG-IMRT, the margins were reduced to 5–7 mm due to higher accuracy of the dose delivery [20]. As mentioned in the introduction, concern has been raised that IGRT and a large reduction in PTV margins might lead to higher rates of local failure. Engels et al. [5] found that IGRT with 3–5 mm margins was associated with a poorer biochemical progression-free survival in patients treated with conformal arc therapy compared to non-IGRT and 6–10 mm margins. Our results are reassuring in that the substantial margin reduction from 2 cm to 5–7 mm did not seem to compromise the biochemical progression-free survival in IG-IMRT patients.

Both 3DCRT to 76 Gy and IG-IMRT to 78 Gy resulted in excellent tumor control. Our results are consistent with the results from other reports of high risk PCa [4, 21]. The percentage of positive biopsies (PPB) was an independent predictor of biochemical progression. It would be easily understood if PPB were a more accurate marker of tumor advancement than the clinical T-stage, which is subject to a high degree of inter-observer variability. Increasing age was associated with a significantly lower risk of biochemical relapse, which might reflect that younger patients might have more aggressive tumors [22].

This study has several limitations. First of all, retrospective assessment of toxicity based on patient records involves a substantial degree of uncertainty. It is also a limitation that 13 of 41 of the relapses in the 3DCRT cohort were based on the initiation of salvage ADT due to PSA progression before relapse according to the Phoenix definition. This might falsely lower the biochemical progression-free survival in the 3DCRT group compared to the IG-IMRT group in which all relapses were based on the Phoenix definition. The two cohorts also differed significantly in baseline characteristics, which might affect the risk of biochemical progression differently in the two groups. Finally, as the median length of follow-up was relatively short in the IG-IMRT group, we were only able to report the 3-year biochemical progression-free survival rates. Thus, our results are somewhat premature, also considering that the large majority of patients received two years of adjuvant ADT.

In this study, the combination of daily image guidance, IMRT and 5–7 mm PTV margins to 78 Gy resulted in significantly reduced GI and GU toxicity compared to 3DCRT to 76 Gy without daily image guidance and with 1–2 cm PTV margins. Importantly, the large reduction in margins did not seem to affect the level of biochemical tumor control in IG-IMRT patients. A major caveat is that this improvement may depend on the details of our implementation, and this should not be taken as a broad confirmation that any implementation of IG-IMRT with reduced margins is always an improvement over 3DCRT with larger margins.