Does TomoDirect 3DCRT represent a suitable option for post-operative whole breast irradiation? A hypothesis-generating pilot study

Adjuvant whole-breast external beam radiation therapy (AWBRT) is an integral part of the current standard multimodality approach for early stage breast cancer[1]. Intensity modulated radiation therapy (IMRT) has been shown to increase intra-target dose homogeneity and to spare organs at risk[2]. Recent studies on breast irradiation[3] have demonstrated that Helical Tomotherapy (HT) provides superior target dose homogeneity and moderate normal tissue sparing compared to conventional treatments. TomoDirect^TM (TD) is a fixed beam treatment mode allowing for planning and delivery at static beams, with the couch moving at a constant rate past a fixed binary multileaf collimator for fluence modulation[4]. A pre-clinical version of the planning system named topotherapy has been investigated in the context of AWBRT, showing adequate target coverage and normal tissue high dose reduction with concomitant PTV inhomogeneous dose in one study[5, 6]. In addition to the IMRT mode (TD-IMRT), it is possible to plan TD in a ‘3D-CRT mode’ (TD-3DCRT), whereas dose penalties and DVH constraints cannot be specified and a preset optimization algorithm determines fluence maps, resulting in an inherently forward plan still allowing for IMRT. TD-3DCRT might be very attractive, since it provides adequate treatment plans and concomitant significantly lower request in terms of planning, calculation and beam-on time resources than TD-IMRT.

Aim of the present study is to investigate TD-3DCRT in the context of AWBRT in terms of dosimetric outcomes of both target and organs at risk and beam-on time and to compare it to TD-IMRT planning and conventional 3DCRT field-in-field technique (FIF). Additionally we defined appropriate patient-class solutions to employ TD-3DCRT in AWBRT, generating the hypotesis that simple anatomic measures could properly drive patient allocation to an appropriate and cost-effective planning modality.

Mean PTV volume was 731 cm³ (range 425–1643 cm³). PTV and OARs cumulative dose volume histograms (DVHs) and isodose distributions for a typical left-sided breast plan are illustrated in Figure1.

The PTV homogeneity of TD-3DCRT plans was significantly better than TD-IMRT (p < 0.01) and FIF (p < 0.001); the PTV conformality was significantly less than FIF (p < 0.05), while no differences were found with respect to TD-IMRT.

The correlation between CLD and ipsilateral lung dose was statistically significant for both V_20Gy and MLD, with Spearman’s ρ = 0.909 (p < 0.0001) and 0.925 (p < 0.0001), respectively (Figure2). Similarly, HD correlated with heart V_25Gy and MHD, with Spearman’s ρ = 0.827 (p < 0.004) and 0.851 (p < 0.002) (Figure2).

The setting of the compensation factor did not significantly affect the dose metrics both for PTV and OARs. Deselecting the normal tissue homogeneity option, increased the volume encompassed by the 95% isodose line and therefore significantly reduced the radiation conformity number from 0.622 to 0.541 (p < 0.05), as increased the D_2cc of approximately 4%.

The mean beam-on time for TD-3DCRT was 295 s (range 219–408 s), significantly shorter than TD-IMRT, with a mean time of 378 s (range 267–475 s, p < 0.001). The use of the low compensation factor produced a mean reduction of beam-on time of approximately 15%.

While TD-IMRT has already been studied, TD-3DCRT still deserves investigation. In the preliminary study of Reynders et al.[5], TD-IMRT showed excellent PTV coverage and OARs sparing compared to conventional 3DCRT and HT. However, a dose inhomogeneity within the PTV (hotspots surpassing 115% of the prescribed dose) was reported, potentially calling for multiple fields with a consequent increase in treatment time. Schubert et al.[6] compared TD-IMRT to conventional 3DCRT: both modalities provided adequate PTV coverage, but TD-IMRT resulted in a significant reduction of high doses to normal tissue.

TD-3DCRT allows intensity modulation with lower resources request in terms of planning and treatment time compared to TD-IMRT. Specifically, TD-3DCRT employs a preset optimization algorithm, only requiring the specification of beam angles, target volumes and dose prescriptions. Consequently, planning and calculation time are consistently lower and comparable to those of conventional 3DCRT. Conversely, beam-on time resulted in a 22% average decrease, which can be further reduced to less than 4 min using the compensation factor low. These considerations make TD-3DCRT very attractive, mainly for radiotherapy departments without conventional Linacs and only equipped with helical tomotherapy units. To evaluate whether this technique might represent a viable solution, we compared 17 TD-3DCRT plans to TD-IMRT and conventional FIF 3DCRT, focusing on PTV and OARs dosimetric endpoints.

The results show that TD-3DCRT provided slightly better PTV coverage and dose homogeneity, as underlined in Figure3, reporting PTV differential DVHs of a sample patient. This can be explained taking into account that in 3DCRT mode, dosimetric constraints are limited to prescription dose in the target as a dose-volume-histogram (DVH) point. Considering PTV average dose and corresponding SDs, TD-3DCRT DVH provided the most reduced width and skew, whilst FIF DVH is skewed towards high dose volumes. Conversely, the symmetry of the PTV differential DVH shows the different PTV homogeneity level achieved by the 3 techniques. Dose to 1% of the PTV was found substantially equal for TD-3DCRT and TD-IMRT, pointing out similar efficacy in limiting maximal dose within the prescription volume.

The dosimetric concern with TD-3DCRT for AWBRT treatments is mainly represented by the radiation dose delivered to the OARs: as shown by our results, both lungs, heart (for left-sided breast cancer patients) and contralateral breast doses are significantly higher than those obtained with TD-IMRT planning. This is particularly evident for ipsilateral lung, if MLD or V_20Gy are considered and for heart if MHD or V_25Gy are taken into account (see Table2 for details). As a consequence, the use of TD-3DCRT in the context of AWBRT might not necessarily be generalized and applied to every typology of patients.

Hence, the definition of specific patient-class solutions, obtained correlating the OARs dose constraints with patient specific anatomic parameters such as the CLD and the HD, might be extremely helpful in creating a subset of patients better suitable for TD-3DCRT, in which the intrinsic spatial relationship between target volume and normal tissues might be able to compensate for the eventual dosimetric slight disadvantage, with no consequent detrimental effects.

Radiation-induced lung injury (RILI) variously occurs (1%-50%) after AWBRT according to different series (prospective or retrospective) and to the heterogeneity of the scoring systems and clinical endpoints employed[12]. If clinically significant radiation pneumonitis (RP) is considered, it may be observed in 1–5% of cases, whereas subclinical RILI (pulmonary function tests worsening/radiological changes) is more frequent[13]. MLD and V_20Gy are reliable dosimetric predictors of RP[13]. Dosimetric threshold on this parameters are used in the planning scenario to decrease toxicity risk.[12]. AWBRT is frequently delivered to prevent from local recurrence good prognosis women; hence it is crucial to adopt safe dosimetric constraints. In our study, we obtained ipsilateral lung V_20Gy = 13.2% (V_20Gymax = 20.3%) and MLD = 6.9 Gy (MLD_max = 10.3 Gy) for TD-3DCRT, significantly higher than for FIF and TD-IMRT. In the different series correlating RP probability and dosimetric predictors, it is rather uncommon to observe an increase in RILI probability for V_20Gy < 20% and MLD < 10 Gy[12]. However, Tsujino et al.[14], evaluating 71 lung cancer patients treated with concurrent chemoradiation, reported 2 cases of Grade 1 RP (according to NCI-CTC v 2.0) for V_20Gy = 15%. The NKI and Duke University series[15, 16], described few cases of RP for MLD < 10 Gy (grade 2 according to SWOG and all grade according to Duke University criteria, respectively). In details, the Duke University study reported a 10% occurrence rate of RP for MLD < 10 Gy, while the NKI study decribed 1 case of grade 2 RP in the breast/lymphoma cohort related to a biological MLD (NTD_mean) of 8–12 Gy.

Hence, we chose 2 strict lung dosimetric parameters (V_20Gy < 15% and MLD < 8 Gy) to allocate patients to TD-3DCRT or TD-IMRT. CLD predicts lung volume involved within the treatment field during AWBRT. A standard tangential approach covers a 2–3 cm CLD measured on a simulation film. Das et al. estimated that 3–3.5 cm CLD represents a lung volume of 15–26%[17].

The risk of symptomatic RP calculated on CLD datasets is < 2% for a CLD range of 1–3 cm and up to 10% for CLD > 4 cm[18]. In our study we correlated V_20Gy, MLD and CLD within a linear fit (Figure2). A dosimetric threshold on V_20Gy = 15% and MLD = 8 Gy corresponds to a CLD cut-off of approximately 3 cm. This CLD value might be a useful clinical tool to predict the opportunity to employ TD-3DCRT (for CLD < 3 cm) or TD-IMRT (for CLD ≥ 3 cm) according to anatomical characteristics to minimize RILI. To further test the safety of TD-3DCRT in terms of lung dose, we selected the 9 patients allocated to TD-3DCRT according to CLD value and analyzed ipsilateral lung V_30Gy and V_40Gy. No differences were found between TD-3DCRT and TD-IMRT, confirming the equivalence also at medium-range lung doses (Table2).

Radiation-induced myocardial damage involves several components (pericardium, valvular and conduction systems, myocardium and coronary arteries).

The dose–response relationship for cardiac risk has been extensively studied, but several uncertainties still remain. The relative seriality model and long-term cardiac mortality clinical data from Scandinavian trials, lead to a threshold dose of 20 Gy for ischaemic disease death[19]. Adjunctively, for conventionally fractionated partial heart irradiation, the same conservative NTCP model-based estimate predicts that V_25Gy < 10% corresponds to a cardiac death probability < 1% at 15 years[19]. Conversely, a study by Carr et al.[20] demonstrated a coronary artery disease relative risk increase in peptic ulcer patients whose cardiac apex exceeded 12 Gy (estimated MHD > 2.6 Gy).

This finding is in line with atomic bomb survivors data where a cardiac mortality increase was documented for 0–4 Sv whole body uniform doses (roughly corresponding to 0–4 Gy, since gamma rays irradiation has a negligible neutron component), with a threshold value of approximately 0.75 Sv[21]. MHD strongly correlates to normal tissue complication probability (NTCP) in terms of cardiac-morality, for left-sided AWBRT[22]. The present study showed a heart V_25Gy range of 1.9–14.3% for TD-3DCRT, higher than for others techniques. Figure2 plots V_25Gy as a function of HD for TD-3DCRT. A linear correlation between V_25Gy and HD was found with a possible 3 cm threshold for HD to properly allocate patients to TD-3DCRT considering cardiac toxicity risk (TD-IMRT for HD ≥ 3 cm). Conversely, MHD did not seem to be a clinically useful screening method for patient allocation. In this sense, a safe dosimetric threshold (MHD < 2.6 Gy) would lead to a HD of roughly 1.5 cm forcing patient planning with TD-IMRT. A permissive dosimetric threshold (MHD < 20 Gy) would allow for a TD-3DCRT planning for the whole cohort. The allocation pattern for all patients according to CLD (>3 cm) and HD (>3 cm) is shown in Figure4. Two patients (12%) presented both CLD and HD > 3 cm, while 5 patients (29%) showed a CLD > 3 cm. Consequently, 10 patients (59%) were potentially allocable to TD-3DCRT. However, 2 issues should be considered to definitively allocate patients: normal tissues dose hotspots external to PTV and contralateral breast dose. TD-3DCRT showed a significantly higher maximum dose outside the PTV (mean value = 8.3% of the prescription dose). These hotspots do not represent a consistent issue, since they involve a very limited volume outside PTV and they do not exceed 10% of the prescribed dose, a threshold for acute skin radiation-induced toxicity during AWBRT[23]. To further reduce dose hotspots in beam entrance and exit regions, one can use multiple beams at different angles or contour ad-hoc planning helping structures, but at the cost of overall planning time increase or eventual higher dose to OARs. Hence we decided to keep planning procedure agile. The dose homogeneity option was always selected, since it minimizes hotspots within normal tissues and its deselection reduced the radiation conformity number without benefit.

Regarding the contralateral breast, TD-3DCRT provided mean D_0.1cc values < 5 Gy (Table2); however it provided higher doses than TD-IMRT, with a contralateral breast D_max range of 0.9–6.7 Gy, similar to conventional 3DCRT. Boice et al. investigated secondary contralateral breast cancer, reporting a relative risk of 1.19 of developing second breast cancer after previous AWBRT (estimated average radiation dose = 2.82 Gy, Dmax 7.1 Gy)[24]. Thus, a dose range between 4–5 Gy to the contralateral breast might be considered sufficiently safe in terms of radiogenic second malignancies and consequently in the present study we focused on a cut-off dose of 5 Gy to definitely allocate patients to TD-3DCRT (contralateral breast dose < 5 Gy) or to replan with TD-IMRT (contralateral breast dose ≥ 5 Gy). Consequently, in our study, the number of patients allocated to TD-3DCRT was finally reduced to 7 (41%).

TD-3DCRT and TD-IMRT are two feasible, reliable and dosimetrically acceptable treatment approach to deliver AWBRT, with optimal PTV coverage and adequate OARs sparing. The lower request of resources in terms of planning, calculation and beam-on time of TD-3DCRT makes this solution very attractive. However some concerns should be raised regarding eventual undue dose delivered to OARs (lungs and heart in particular) if TD-3DCRT is applied to a general patient population. A correct patients clusterization according to simple quantitative anatomic measures (such as HD and CLD), would help to correctly allocate patients to the appropriate treatment strategy in terms of target coverage, but also of normal tissue sparing.