Radiotherapy of large target volumes in Hodgkin's lymphoma: normal tissue sparing capability of forward IMRT versus conventional techniques

Radiation treatment and antiblastic chemotherapy of Hodgkin's lymphoma (HL) is a proven curative therapeutic strategy capable of curing the vast majority of patients. Radiotherapy in Hodgkin's lymphoma is very often characterized by fields encompassing different body sites. The great variability of thickness and density in the irradiated tissues makes it difficult to achieve a homogeneous distribution of the dose. Moreover, the low average age of HL patients, in the cases in which a large volume needs to be irradiated, makes these patients' population particularly at risk of developing late side effects and secondary neoplasms. The irradiation of the thyroid region, for instance, induces a 50% risk of developing hypothyroidism and a 20% risk to develop thyroid nodules [1‐3]. Radiation dose and irradiated volume of the thyroid gland correlate with the incidence of hypothyroidism. In particular, the volume of gland receiving a dose greater than 30 Gy has been shown to significantly impact on the TSH peak [4]. Irradiated volume at given dose levels can be also related to late cardiac and pulmonary toxicity [5, 6]. For example, the risk of grade 3 late lung toxicity has been found to be 38% or 4% depending on whether the volume receiving 25 Gy is larger or smaller than 30% respectively [7]. Several papers have reported an increased risk of breast cancer in girls and young women among HL patients: breast cancer represents 6.3 to 9% of all secondary cancers occurring after HL treatment [8]. Higher radiation doses might increase the risk of developing breast cancer. Tailoring radiotherapy to eliminate as much breast tissue as possible from the radiation field may reduce this risk [9]. All these issues have to be carefully considered by the radiation oncologists approaching the therapeutic strategy in HL while the medical physicists have to make every possible effort to optimize treatment plans.

Perhaps, due to the low doses used in the treatment of this disease, only in recent years some efforts have been made to improve dose distribution in HL treatment plans. Several delivery techniques, such as intensity modulated radiation therapy (IMRT) techniques with or without inverse planning optimization and even three- dimensional proton radiotherapy, have been proposed in the literature [10‐16]. All these techniques aim at achieving better homogeneity in target dose distribution and dose reduction to critical structures. However, there have been some discussions on IMRT techniques for large planning target volumes (PTVs) and their actual implementation due either to field size restrictions using dynamic multileaf collimators [17, 18] or to the greater volume of normal tissue receiving low-to-moderate radiation doses and its related late radiation effects [19].

In this work we define and quantify the dosimetric advantages of a forward planned intensity modulated technique (FPIMRT) via segmented fields [20] for selected Hodgkin's lymphoma patients for whom large field irradiation is required. To this purpose we have simulated ten consecutive HL patients undergoing post chemotherapy involved field radiotherapy with PTV larger than 900 cm³. For each patient two treatment plans, a conventional parallel-opposed field plan and a FPIMRT plan, were retrospectively generated. Dose homogeneity in the target and normal tissue sparing capability were the main focus of our analysis.

Ten patients with Hodgkin's disease who had received post chemotherapy radiotherapy at the Department of Radiotherapy of the University "Federico II" of Naples were retrospectively considered for the study. These patients had stage II disease requiring a large volume of irradiation. They represent about 6% of the patients treated in the last 9 years at our department. Patients and disease characteristics are shown in table 1. Mean age was 25.6 years (95% CI, 18.7-32.5). In all patients a continuous CT-scan was performed in supine position using vacuum locked mattress with the arms up above the head. Scans were acquired using 5-mm slices of a multislice scanner with the craniocaudal limits, generally 4 cm behind the target region.

CT images were electronically transferred to the Focal Ease 4.2 CT Simulation software (Computerized Medical System, Inc., St Louis, MO) for the contouring of target and critical organs (lung, spinal cord, heart, thyroid and, in women, breast). Target volumes and organs at risk were delineated by the same radiation oncologist (R.P.) and checked by a senior radiologist (L.Ca.).

Clinical target volume (CTV) included the nodal sites involved at the time of diagnosis. The nodal sites were delineated according to the modalities in use for three dimensional conformal radiotherapy (3D-CRT) in solid tumors. Namely, for the neck we referred to the internationally accepted guidelines of Gregoire et al. [21], to Mountain and Dresler [22] for the mediastinum, and to Dijkema et al. [23] for supraclavear and axillary nodes. Planning target volume (PTV) included CTV plus a 10 mm margin. For this study, we considered patients with a target volume larger than 900 cm³ (all PTV volumes are reported in table 1).

Treatment planning was done by a 3-D planning system (XiO 4.4, Computerized Medical System, Inc., St Louis, MO). Two new treatment plans were on purpose generated for each patient: conventional anterior-posterior and posterior-anterior (AP-PA) plan and FPIMRT plan. Both plans were simulated using 6 MV X-rays with a dose rate of 200 MU/min, from Siemens Primus (Siemens Medical Systems, Erlanger, De) linear accelerator equipped with 29 pairs of double-focused multileaf collimator (MLC). A total dose of 30 Gy in 20 daily fractions of 1.5 Gy was planned. The same physicist performed all treatment plans. For both techniques, treatment plans were optimized to ensure, when possible, that 95% of the prescription dose was delivered at least to 95% of the PTV and, at the same time, with a maximum dose less than 120%.

The dose distribution was calculated using the Xio Multigrid Superposition algorithm [24] appropriate in the presence of heterogeneous tissues.

Radiation treatment of Hodgkin's lymphoma is an efficient therapeutic modality that, coupled with antiblastic chemotherapy, can cure the large majority of patients [30, 31]. Despite substantial advances in radiation treatment techniques in many areas, radiotherapy for HL is still delivered in a conventional way in most radiotherapy departments, and dose gradients that do not perfectly comply with general RT guidelines recommendations are often accepted.

Alternative delivery techniques with different complexity levels aimed at achieving better target coverage and critical structures sparing have been recently proposed.

A sliding window mantle technique [12], using dynamic MLC (dMLC) and electronic tissue compensation, succeeded in obtaining a better and more homogeneous target coverage in comparison with the conventional plan. However, the monitor unit number was increased by a factor of 3 in dMLC plan. Some investigators have proposed the use of IMRT for HL fields. Goodman et al. [13] used IMRT to irradiate lymphoma patients selected on the basis of either a large mediastinal treatment volume or because particularly at risk (reirradiation or previous antracyclin based treatment). The latter showed an improved target coverage and an amelioration in the pulmonary toxicity profile. Girinsky et al. [14] showed that, for mediastinal HL masses, IMRT achieves a better dose conformation and PTV coverage compared to 3D-CRT. Moreover, the heart, coronary arteries, esophagus, and spinal cord were more protected with IMRT plan, the only drawback being a greater volume of tissue receiving low doses compared to the conventional plan. Indeed the median dose delivered to the body increased seven folds. As the authors pointed out, this can be of concern in relation to the young age and long life expectation of HL patient population. Furthermore, IMRT technique becomes particularly complex in those cases in which large volumes have to be covered. Large IMRT fields cannot usually be implemented using available linacs because of issues related to MLC design [17]. In addition, because of the considerable cost and requirement of human resources, there are still many centers that have no adequate funds to implement this technique.

A simple forward planned IMRT technique has been suggested, in which dose conformation is obtained by combining MLC AP - PA fields and segments, with simple beam weighting modulation [10, 11]. The authors describe better dose homogeneity and only assume a reduction in complication rate compared to conventional methods. In our clinical practice the FPIMRT is currently the standard technique for HL radiation treatment, regardless of the target dimensions.

The present report expands on the potential of the FPIMRT technique and extends the complexity of the analysis in order to evaluate and quantify the possible advantage of this technique vs. the conventional one in the case of large treatment fields in Hodgkin's lymphoma. Starting from an accurate and reproducible delineation of the target volume and of the OARs, the comparison was made considering normal tissue sparing capabilities. Dose volume constrains and NTCPs were the main focus of the evaluation. Indeed, in the past few years, the FPIMRT has been utilized to improve dosimetry in radiation therapy planning, and its general advantage on PTV coverage and homogeneity has been well documented. However, to our best knowledge, the advantage on normal tissue sparing in HL is only hypothesized and not quantified.

In addition, in this work we propose a reproducible way of drawing the target in HL patients following the nodal delineation suggested by other authors [21‐23] for 3D conformal radiotherapy in solid tumors. Indeed, since the great variability in target definition represents a critical issue in the evaluation of different techniques [32], some standardization is needed.

As regards dose homogeneity and target coverage, we obtained good results with the FPIMRT technique compared to the conventional AP-PA treatment, being the dosimetric parameters for PTV significantly better with the FPIMRT plan. Indeed, adding segments in the right positions with appropriate weights allows to avoid, at the same time, hot and cold spot regions characteristic of the AP-PA treatment. As shown in table 3, all mean dosimetric parameters for PTV are similar to those obtained by other authors with full IMRT on large PTV [13]. We could not make a direct comparison on our patients since in our centre we don't have the suitable technology to perform IMRT for large treatment fields.

Despite the simplicity of the FPIMRT technique and the large PTVs considered, the obtained results were encouraging when we also consider doses to critical organs and the related toxicity rates. We found that the FPIMRT technique allows a reduction of normal tissue complication probability in all critical structures other than the heart for which both the NTCPs and the dosimetric parameters resulted comparable.

The appropriate parameters to be used to describe the probability of pulmonary toxicity are a matter of debate, and different predictive parameters have been proposed in literature [6, 33, 34] including the mean lung dose and the V10-V30. Considering our results, we can see that while the mean lung dose and V20 were similar for the two different techniques, V30 was significantly reduced with the FPIMRT plan. This result could indicate a lower pulmonary toxicity since radiation pneumonities rates seem to be correlated with a reduction in higher dose volume rather than with a reduction in lower dose volume [33]. If we compare our results for lungs with those obtained with IMRT on large volumes [13], we obtain a somewhat higher mean lung dose whereas, with the IMRT, V20 was greater.

The FPIMRT technique has also the advantage, when compared to the conventional one, of decreasing the maximum dose to the spinal cord (fig. 3d). However, no change was found for breast irradiation.

The results of our analysis are particularly striking when considering the thyroid gland: all dosimetric parameters and NTCP improved. Indeed, for all patients, V30 (figure 4c) resulted greatly reduced with a consequent lowering of hypothyroidism risk [4].

An indirect comparison of FPIMRT with full IMRT optimization suggests that, for smaller target volumes, full IMRT allows a better sparing of the heart and the coronary arteries from the high dose region [14]. Nevertheless, when larger target volumes are considered, because the advantages of full IMRT in heart sparing decrease [13] and the associated workload increases, this more sophisticated technique doesn't seem worthwhile. Another aspect that must be considered, especially in young patients, is the risk of induction of secondary malignancies which may result from larger low dose tissue volumes with IMRT [16].

As a whole, when considering target coverage improvement, OAR sparing capabilities, the ease of execution and delivery time, the use of the FPIMRT technique shows not only a definite improved performance when compared to the conventional AP-PA technique, but also represents a valid alternative when more sophisticated techniques are not available.