Comparison of conventional and CT-based planning for intracavitary brachytherapy for cervical cancer: target volume coverage and organs at risk doses

Intracavitary brachytherapy (ICBT) with external radiotherapy (ERT) is an essential component of cervical cancer management and has a high therapeutic index by delivering a high dose to the primary cervical lesion and lower doses to adjacent organs, resulting in increased local control and survival without increased in toxicity [1–4]. However the doses delivered to tumor and normal tissues from ICBT are difficult to quantify accurately in conventional brachytherapy (BRT) planning. To ensure consistency in the reporting of ICBT applications in cervical cancer, the International Commission on Radiation Units and Measurement (ICRU) recommended a number of parameters for doses and volumes to be considered. These include points A and B, representing the doses in the parametria and the pelvic wall, and the rectal and bladder points representing the organs at risk (OARs), respectively [5]. Physicians have used these reference point doses to report treatment intensity and to estimate the maximal dose to normal tissues, which can predict late complications. However, the conventional plan in treating cervical cancer may not correspond with the individual extent of the tumor. This may result in either undercoverage of the tumor or overdosage of the surrounding normal tissue.

A modern approach in treatment planning for cervical carcinoma is based on computed tomography (CT) sections and on a 3D dose distribution. This allows better assessment of dose distributions in different volumes, such as the gross tumor volume (GTV), clinical target volume (CTV), and OARs (rectum, bladder, and intestines). Ling et al. published the first report describing the volumetric dose distributions from ICBT [6]. In 2004, guidelines were published for proposing image-based BRT for cervical cancer [2]. However, the results of the first preliminary studies indicated that a great deal can be learned from volumetric analysis of ICBT dose distributions [7–9]. Furthermore, the actual doses delivered to the tumor, bladder, and rectum during ICBT do not correlate well with those estimated from ICRU reference-dose calculations, demonstrating that the point A dose in conventional plans overestimates the target volume dose coverage and underestimates the OAR doses determined by CT plans [10–12].

Although conventional treatment planning has generally yielded high tumor control rates, with a low frequency of major complications, a more accurate understanding of the radiation doses delivered during ICBT may lead to improved treatment outcomes. In an attempt to solve some of the problems that have limited the use of volumetric analysis of ICBT dose distributions and to achieve a better understanding of the treatments, we compared two treatment planning methods based on orthogonal radiographs (conventional plan) and CT sections (CT plan). The comparison was based on point doses defined by the ICRU and dose volume histograms (DVHs) from 3D planning.

In the current study, we assessed the conventional BRT plan based on ICRU reference points and the CT-based BRT plan in patients with cervical cancer. We clearly demonstrated that tumor volume coverage was inadequate in the conventional plan compared to the CT-plan, and was inversely related with the volume of the target and the extension of tumor. With the conventional plan, the ICRU rectum and bladder point doses underestimated the actual rectum and bladder doses obtained from the CT-plan. Additionally, we demonstrated that more precise analysis of the dose received by certain volume of OARs can be accomplished by utilizing the DVHs on CT-plans, which may be of critical importance in regard to normal tissue tolerance limits.

After publication of ICRU 38 report, ICRU reference points for tumors, and reference dose points for bladder and rectum were used for defining the doses in conventional plans. But calculation of doses with these fixed reference points relative to applicators has certain limitations. The conventional plan with the point-A dose calculation relies on reference points on orthogonal films, not tumor volumes defined on CT, which may cause underestimation of tumor doses. Likewise, the calculation of rectum and bladder doses made with ICRU reference points, not with rectum and bladder volumes, may not reflect the actual organ doses. In addition, sigmoid colon and small bowel in the pelvis may be in close proximity to the BRT sources during application, and the doses to these organs should also be assessed. Since the ICRU did not define standard points for the sigmoid colon and small bowel, it is not possible to evaluate doses to these organs with conventional plans. To overcome such problems, CT-guided 3D BRT treatment planning has been used successfully for customizing the dose distribution according to tumor extent and providing detailed dose-volume information on the target volumes and surrounding tissues [12, 17–21].

Some investigators have reported that the point A-dose in the conventional plan overestimates the target volume dose coverage [10–12]. In addition, more advanced tumor stages and larger target volumes receive less coverage with the prescribed dose, which may result in poor local control [12, 22]. Datta et al. demonstrated that the percentage of tumor encompassed within the point-A dose envelope ranged from 60.8% to 100%, and this percentage depended on the tumor volume at the time of ICBT [18]. In the current study, we demonstrated that the mean percentage of GTV and CTV encompassed within the point-A 7 Gy isodose level was 93.1% (74.4%–100%) and 88.2% (58.8%–100%), respectively. Inadequate tumor coverage could significantly influence the treatment outcome in patients, especially in those who have partial regression of tumors with gross residual tumor after ERT. Thus, tumors with larger volumes at ICBT were more likely to have portions outside the 7 Gy prescribed isodose line (Figures 2 and 3). Initially, Kim et al. demonstrated that the CT-plan would be beneficial in patients with large CTVs, which could not be fully encompassed by the 100% isodose line [12]. In the current study, the GTV and CTV were larger in group 2 than in group 1; therefore, the CT-plan would be most beneficial in group 2. Although the isodose matrix volumes did not differ between the two groups with the conventional plan, these volumes were higher in group 2 with the CT-plan (Table 2), which may cause a significant incremental dose to the neighboring tissues, mainly the bladder and sigmoid colon (Table 3).

Although tumor shrinkage before BRT applications may take place after ERT, the initial tumor stage, which reflects the tumor extension, may negatively impact tumor coverage [1, 22, 23]. Kim et al. demonstrated that GTV but not CTV increased with advanced stages [23]. They also found that the percentages of the GTV encompassed by the 6 Gy isodose line were 98.5%, 89.5%, 79.5%, and 59.5% for stages IB₁, IB₂, IIB, and IIIB, respectively. In our study, the GTV and CTV appeared to increase with more advanced clinical stages. Meanwhile, tumor coverage within the 7 Gy isodose line diminished with more advanced clinical stages (Table 3). Therefore, a higher stage of tumor received less coverage by the prescribed point-A dose because of extension to the parametria and/or vagina.

For evaluating the maximum doses to OARs, the dose to a clinically significant volume is used; that clinically significant volume can be defined as the volume exposed to a minimum dose in the part of the OAR that receives the highest dose. The size of this volume can be absolute (e.g., 1, 2, 5, or 10 cc) or relative (e.g., 1%, 2%, 5%, or 10% of the contoured OAR). Several investigators have compared the dose volume based on either the exterior organ contour or only the organ wall, for the bladder and rectum [8, 24, 25]. To evaluate organ wall dose correctly, the volume of 2.0 cc is considered, because the D2 computed for the external contour are almost the same as the D2 to the organ wall. Also, this 2.0 cc volume of tissue in the highest dose region is probably more clinically relevant. Although the difference between the DVHs increases greatly for volumes larger than 2.0 cc, we also chose the dose of a 5-cc volume (D5), because this volume was previously reported as the minimal volume required for fistula formation [7, 8].

The rectum and bladder doses were found to be greater than the corresponding ICRU reference doses [7, 8, 12, 18, 26]. In these other studies, the true bladder and rectum doses were 1.5–2.5 times greater than the corresponding ICRU reference point doses. Pellioski et al. compared the minimal doses delivered to 2 cc of the bladder and rectum (D_BV2 and D_RV2) and found that ICRU bladder reference point dose was significantly lower than the D_BV2, but the ICRU rectum reference point dose was not significantly different from the D_RV2 [26]. Our study indicated that the maximum rectum and bladder D2 values were 1.66 and 1.51 times greater than the ICRU reference rectum and bladder doses, respectively. We also found that the maximum rectum and bladder D5 values were 1.42 and 1.28 times greater than the ICRU reference rectum and bladder doses in CT plan. When we evaluated the difference between the ICRU rectum and bladder doses and corresponding D2 and D5 values, the differences between the ICRU bladder point dose and D2 and D5 bladder doses were significantly higher in group 2 than in group 1; however the difference in rectal doses did not differ significantly (Table 5).

Since the sigmoid colon and small bowel in the pelvis are close to the radiation source during ICBT, doses received by these organs should also be assessed. The ICRU defined the reference points for bladder and rectum, the initial dose calculations for these organs were performed during the conventional plan. In addition, the doses to the sigmoid colon and small bowel can be evaluated with the CT-plan using DVHs. Al-Booz et al. pointed out that the sigmoid colon received doses in excess of 70% of the intended point-A dose during BRT [27]. Kim et al. demonstrated that that the sigmoid colon received the highest mean D2 when compared to the rectum and small bowel [28]. Their study revealed that with the prescribed dose of 600 cGy, the sigmoid colon received the highest mean D2 (408 cGy) followed by the small bowel (379 cGy), and rectum (373 cGy). In our study, we clearly demonstrated that the small bowel D2 was higher than the sigmoid colon D2 (6.8 Gy and 6.5 Gy, respectively). We also found that the sigmoid colon D2 and D5 values were significantly higher with larger CTVs (Table 4). The small bowel D2 values were higher in group 2 than in group 1, and this difference was almost statistically significant (P = 0.07).

The results of our study demonstrate that CT-guided BRT planning is superior to conventional point A planning in terms of both conformity of target coverage and evaluation of OARs, including the sigmoid colon, bowel, bladder, and rectum. Although this superiority was clear for small CTVs, for large CTVs both the conventional and CTV plans had the drawbacks of inadequate target coverage and/or excessive radiation doses to normal organs. To ascertain the potential benefit of treatment outcomes, such as tumor control probability and morbidity, ICR with image-guided 3D planning will be pursued and correlated with the dose-volume parameters.