Benefits of proton therapy across various subsites are well established and its role in low income countries like India is well justifiable [14]. As a first modern proton therapy centre in South East Asia, we treated more than 500 patients over past two years and CNS tumours constituted maximum bulk of it. We have reported our preliminary experience in children and young adults [15, 16].
Indications of CSI and technical evolution in treatment delivery over last few decades we discussed earlier. Delivering radiation to such large volume target is complex procedure and teletherapy machine limitations for such long treatment, warrants use of multiple fields with field junctions. Field junctions, location of vital OARs (Heart, Lungs, Midline mucosa and bowel bag) in close proximity to target and longer survivorship of patients receiving CSI mandates use of highly precise radiation therapy technique.
In addition, most of these patients warranting CSI are of paediatric age group followed by young adults; due all these facts CSI planning and delivery is considered as critical procedure. Over last six decades, CSI techniques evolved tremendously and with availability of proton therapy, it is current standard of care whenever available. In paediatric patients as per current standard recommendations (SIOP and COG) target volume should include whole vertebra so as to avoid or minimize risk of radiation induced skeletal growth abnormalities [10].
Proton beam therapy with pencil beam scanning technique has significantly better dose distribution and OAR sparing when compared with modern intensity modulated radiotherapy and passive scattering proton beam therapy.
In this work, we are analysing acute toxicities of CSI using PBT in both young adult and paediatric patients and their comparison with existing results. Like discussed in various dosimetric studies, proton therapy with no exit dose limits radiation to anteriorly placed OARs due to absence of exit dose [8, 17]. Dose distribution in sagittal and axial sections with OAR sparing achieved for adult patients, paediatric patients with standard approach and our novel approach are shown in Fig. 2. Figure 2 also shows that due to physical properties of proton with no exit dose, it allows limitation of radiation doses to anteriorly placed OARs. In our study population, radiation doses received by all OARs are significantly less when compared with published photon literature and are similar or on lower side compared to proton literature. It is important to achieve good coverage in cribriform plate region as it is one of the common sites of disease recurrence and its proximity to lens and cornea poses difficulty in radiation planning. To achieve better coverage in cribriform region and limit lens, optic apparatus radiation doses we used two posterior oblique fields for brain which in comparison to standard bilateral fields provided more than 98% of prescribed dose coverage without increasing lens dose [11].
Mean dose of bilateral parotids was again significantly less, while average dose received by lungs, kidney and bowel was less than 0.1%. Liver, heart and gonads did not receive any radiation dose. We did subgroup analysis by dividing these patients in paediatric (less than 14 years) and adult age group (> / = 14 years), as all our patients less than 14 years received whole vertebra CSI. In adult patients, radiation dose to all OARs was less than 0.1% of prescribed dose as shown in Table 2. For paediatric patients in whom whole vertebral CSI was planned, we modified standard contouring guidelines as in this group of patients with standard contouring guidelines, vital OARs such as midline mucosa, oesophagus, thyroid gland and bowel bag are partially in PTV volume (as shown in sagittal Fig. 2IIA, and axial section Fig. 2IIB). And this causes either significant dose delivery to the OARs up to 100% of prescribed dose or significant underdosing in the target volume. We mitigated this with our modified contouring method in paediatric patients receiving proton CSI, and like discussed above. We instead of accepting under-dosing or sharp dose gradient in vertebral body region, accepted uniform low dose and verified vertebral body coverage while radiation plan robustness (3 mm translational shifts and 3.5% range uncertainties) was done for CTV including spinal canal and nerve roots. This novel approach, allowed better OAR sparing with no or minimal target underdosing, no dose gradient in vertebral body region.
Dosimetric benefit of proton therapy for CSI is translated in reduction radiation dose to OARs which in turn reduces acute toxicities and improves compliance to treatment. Acute toxicities in our cohort were remarkably less. In overall population most common acute toxicity was dermatitis, 50% patient had grade one dermatitis while remaining 50% developed grade 2 dermatitis and no patient developed grade 3 or more dermatitis. All patients developed dermatitis in scalp region near boost volume and no patient developed grade 2 or more dermatitis in back region near spinal target volume.
Out of these 40 patients, only 8 (20%) developed grade 1 mucositis while 2 (5%) patients developed grade 2 mucositis and one (2.5%) developed grade 3 mucositis. In subgroup analysis, only one patient aged more than 14 years developed grade 1 mucositis, no one developed grade 2 or more mucositis. In patients aged less than 14 years, 22% developed grade 1 mucositis while only two patients (6%) developed grade 2 mucositis and one (3%) developed grade 3 mucositis. Only patient who developed grade 3 mucositis was first paediatric patient receiving whole vertebral body CSI using PBT and contouring was done as per the standard SIOP guidelines (Fig. 2IIA, B). In addition, patient was receiving concurrent weekly injection vincristine. After 5 fractions he developed grade 3 esophagitis, required treatment break, oral opioid analgesics and intravenous fluid administration with supportive care. For him plan adaption was done with partial underdosing of vertebral body in PRV region to limit OAR doses as per Tasson et al. [18]. After this we reviewed existing literature and mitigated this issue with modified approach as discussed above, and with this approach no paediatric patient developed grade 3 or more toxicities.
Haematological toxicities were noticed in 60% of all patients, 15% patients had grade 1 neutropenia, 7.5% had grade 2 neutropenia and only four (10%) patients developed grade 3 neutropenia. Again grade 3 neutropenia was seen only in paediatric patients either receiving chemotherapy or re-irradiation. Only 3 patients developed grade 1 thrombocytopenia, all were receiving concurrent chemotherapy and of these 2 were from paediatric age group. Four patients had decrease in haemoglobin, 3 had grade 1 decrease while one had grade 2; none developed grade 3 anaemia. Gastrointestinal toxicities were noted in 10 patients. No patient in this cohort developed grade 2 or more weight loss during treatment and median weight loss was 4.8% of baseline weight, 3 patients lost weight more than 10% of their baseline, all were paediatric age group. No patient developed grade 2 or more anorexia, nausea vomiting during treatment, grade 1 anorexia was noted in 7 (23.3%) patients while 6 (20%) had grade 1 nausea vomiting.
Hospital admission was required in five patients, all were age less than 10 years, reason for hospital admission was either supportive care, symptomatic treatment or management of neutropenia or VP shunt management. Average hospital stay as we mentioned above was 2 days and this admission during treatment did not prolong overall treatment time for these patients. Three patients, 2 paediatric and one adult patient underwent plan adaptation during treatment, trigger for adaptation was daily CBCT in two patients while one had clinical indication. Reason for plan adaptation were either set up errors due to change in patient characteristics (67%) or acute mucositis (33%).
At median follow up of 12 months, of these 40 patients, 83.3% patients are alive and of them 92% were alive with no evidence of disease (Fig. 3A). We did subgroup analysis, as this cohort had different diagnosis which are strong determinant of their treatment outcomes. We divided patients according to diagnosis in three groups, group one included patients with medulloblastoma, group two included recurrent ependymoma patients receiving CSI as re-irradiation and others were included in third group.
In medulloblastoma group, 19 out of 20 patients are alive with overall survival of 95% at median follow up of 12 months. In recurrent ependymoma group OS was 71.4% while in third group OS was 76.9% (Fig. 3B). When we compared our toxicity and survival outcomes with available proton publications, our patients had comparable or slightly better toxicity profile and similar survival outcomes. Barney et al. in audit of 50 adult CSI patients treated with proton therapy with similar median CSI dose at median follow up of 20 months, noticed median weight loss of 1.6%, grade 1 nausea/vomiting in 46%, grade 2 nausea in 20% and 10% patient had grade 2 or more anorexia, while four patients had grade 3 or more cytopenia [19]. In comparison in our cohort of adult patients, median weight loss was marginally less (1.6% vs. 1.38%), while none had nausea/vomiting, anorexia and cytopenia.
McGovern et al. in their experience of 14 paediatric patients received CSI using PBT, documented grade 3 or more haematological toxicities (Neutropenia, thrombocytopenia and or anaemia) in 50% of patients, while one developed sepsis during the treatment [20]. In comparison, in our cohort only one (4.5%) patient developed grade 3 neutropenia. Considering different histopathologies in both groups survival outcomes are not comparable.
In addition to this toxicity profile, we expect significant risk reduction in late toxicities and risk of second malignant neoplasm in both adult and paediatric patients treated with proton therapy as predicted by various models [21,22,23], especially in medulloblastoma group where long survivorship is reported [24]. Superior toxicity profile and compliance to treatment in both adult as well as paediatric patients can be attributed to holistic care for each patient, combined team efforts of medical oncologist, dietician, physiotherapist and psychotherapist along with treating physician. We are following standard practice of regular weekly reviews, starting nutritional supplements at the start of treatment and low threshold for intervention. In addition pencil beam scanning technique with daily on board imaging and regular quality assurance CT scan imaging ensured accurate treatment delivery.
Limitations of this study includes a relatively short follow up, mixed age groups and histologies included. We would like to continue follow-up of these patients, document quality of life, late toxicity outcomes and survival outcomes for this cohort.