Journal of Cancer Research and Therapeutics Medknow Publications on behalf of the Association of Radiation Oncologists of India (AROI) ISSN: 0973-1482 EISSN: 1998-4138 Vol. 7, Num. 1, 2011, pp. 58-63

Journal of Cancer Research and Therapeutics, Vol. 7, No. 1, January-March, 2011, pp. 58-63

Original Article

Intensity-modulated radiation to spare neural stem cells in brain tumors: A computational platform for evaluation of physical and biological dose metrics

Arun Jaganathan, Meena Tiwari, Rahul Phansekar, Rajkumar Panta, Nagraj Huilgol

Department of Advanced Centre for Radiation Oncology, Dr. Balabhai Nanavati Hospital, Mumbai - 400 056, Maharashtra, India

Correspondence Address: Arun Jaganathan, Advance Centre for Radiation Oncology, Dr. Balabhai Nanavati Hospital, S. V. Road, Vile Parle West, Mumbai - 400 056, Maharashtra, India, arunradphy@gmail.com

PMID: 21546744

DOI: 10.4103/0973-1482.80463

Abstract

Keywords: EUD, intensity-modulated radiotherapy, Matlab, neural stem cell, TCP

Introduction

In recent years, the concept of neural stem cell (NSC) has ignited a great deal of interest largely due to potential clinical benefit associated with sparing stem cells during irradiation. It is now known that the human brain contains regions of mitotically active cells which retain the ability to divide and differentiate along either neural or glial cell lines throughout life. These neural stem cells are located in two specific areas of the brain: the subgranular zone within the dentate gyrus (part of the hippocampus), ^[1],[2] which allows permanent and irreparable damage during radiation, and the subventricular zone adjacent to the lateral aspect of the temporal horn and the occipital trigone region of the lateral ventricles that have capacity to regenerate to some extent from the effects of ionizing radiation. ^[3],[4],[5] These cells are capable of increasing their mitotic rate under the influence of appropriate stimuli (e.g., brain trauma, stroke, radiation exposure, etc.), and can migrate through the brain to damaged areas and repopulate areas of cortical neuronal loss or white matter damage. ^[6],[7] They are also involved in replacing the neurons that are lost as a result of neurodegenerative disorders, and are important in learning. There is lack of adequate survival statistics or models to extrapolate the hypothetical clinical benefit by sparing NSC.

Extensive laboratory data published recently suggest significant impact of radiation on brain tissue leading to altered cognitive function. ^[8] However, there is considerable uncertainty about the development of these changes. New in vitro and in vivo approaches have provided the means by which new mechanistic insights can be gained relevant to the topic. Irradiation of brain leads to acute and late sequelae resulting in neurological morbidity. Acute effects of radiation to brain mainly manifests clinically as nausea and/or vomiting, irrespective of the site of radiation. Marsh et al. have reported that it is dosimetrically feasible to spare NSC using helical tomotherapy with 65.8% reduction in biologically equivalent dose (BED) _(α/β=10Gy) for the NSC compartment in the prophylactic cranial irradiation (PCI) plans and a 70.8% reduction in the whole-brain radiotherapy (WBRT) plans. ^[9] Preservation of the NSC compartments during the administration of PCI should result in maintenance of the ability of the brain to repair the damage generated by cranial irradiation and help preserve neurocognitive function. Barani et al. have shown that it is possible to identify and dosimetrically reduce dose to these regions using intensity-modulated radiotherapy (IMRT) while treating a patient using treatment schedules applicable to WBRT and a primary high-grade glioma. ^[10]

Materials and Methods

This is a retrospective analysis of patients with grade III and IV gliomas of brain who underwent postoperative radiation therapy using IMRT technique and 6MV photon energy. The radiation dose schedule included 50 Gy delivered in 25 fractions with 200 cGy daily fraction size delivered over 5 weeks for phase 1. A boost of 10 Gy was delivered for phase 2, which was not included in the present dosimetric study. Patients were immobilized using thermoplastic mold in treatment position before computed tomography (CT) and magnetic resonance imaging (MRI). CT and MRI images thus obtained were fused on Tomocon (Tomocon 3.0.13, Slovak Republic) and contouring of NSC done by the radiologist, as shown in [Figure - 1]. Radiation oncologists assigned constraints to various sites like eye, lens, optic nerve, and NSC. Levels of constraints assigned to various subvolumes are shown in [Figure - 2].

All the plans were carefully designed on precise planning system using aperture-based inverse planning algorithm (ELEKTA Ltd, West Sussex, UK). An attempt was made to reduce the number of segments per plan conceptually to 80 segments with maximum of seven beams. Reduction of segments was attempted to reduce time without compromising plan quality. Plans were generated for all patients with and without constraints to NSC. The session of IMRT planning was grouped as Group 1, with NSC as a constraint in optimization process, and Group 2, without including NSC as constraint in IMRT segmentation. The goal of grouping group 1 and group 2 for the current study was to demonstrate the onset of acute toxicity of NSC in terms of BED _{(α/β = 10.0 Gy)} and _{(α/β = 2.0 Gy)} ^[11] for late-responding tissues of the brain.

Plan analysis

Plans in Group 1 and 2 were compared against multiple indices bridged with functional database of computational model. This model functions by using Matlab 7.8.0.347 Version, Microsoft office Excel 2007, and Statistical Analysis Software SAS (Cary, NC) to assess the physical dose metrics as RTOG indices and biological dose metrics such as BED, generalized equivalent uniform dose (EUD), tumor control probability (TCP), NTCP, and physical dose volume histogram (DVH) vs BED-DVH analysis.

The plan quality was deemed acceptable if SD of all the above indices calculated individually were within 0.3% against Group 2 plans. For DVH as an input function, the plan in Group 1 would undergo reiteration till the acceptability criteria are met.

The priority of constraint to NSC is assigned as "x" magnitude varies from +x to -x values as per the model output log files and get into reiteration of plans in Group 1 eventually are within acceptable limits, as shown in flowchart of NSC Model from [Figure - 3].

Physical and biological dose metrics

(a) Physical indices--Radiation Therapy Oncology Group criteria:

To rank the treatment plans, Radiation Therapy Oncology Group (RTOG) proposed certain criteria based on reference isodose (V _RI ) and the target volume (TV) of the treatment plan.

Quality of coverage _RTOG = I min/RI [Index should be 90% coverage] (1)

Where, I min - minimum isodose around the target and RI - reference isodose.

Homogeneity index _RTOG = Imax/RI [Index should be ≤2] (2)

Where, I max - maximum isodose in the target and RI - reference isodose.

Conformity index _RTOG =V_RI/TV [Index should lie between 0.9 to 2]

Where, V _RI - reference isodose volume and TV - target volume.

To evaluate precisely and compensate for the defects in above indices, van′t Riet et al. proposed an index called conformation number (CN). ^[12] Calculation of this CN simultaneously takes into account irradiation of the TV and irradiation of healthy tissues.

CN = TV_RI/TV x TV_RI/V_RI [Index should lie between 0 to 1 (4)

Where, CN - conformation number, TVRI - target volume covered by the reference isodose, TV - target volume, and VRI - volume of the reference isodose.

(b) Biological indices

Niemierko phenomenological model of EUD is able to address tumors and Organ at Risk (OAR). ^[13] EUD is normalized at 2 Gy per fraction and combines the effects of both sensitivity to fractionation and volume effects of target or OAR.

′a′ is a unitless model parameter that is specific to the tumor (<0) and OAR (>0) and v _i is unitless and represents the i′th partial volume receiving dose D _i in Gy. Tissue-specific parameter is "a =-8" for brain tumors. ^[14]

To calculate the TCP, the EUD is substituted in the following equation:

The TCD ₅₀ is the tumor dose to control 50% of the tumors when the tumor is homogeneously irradiated. radiated. γ₅₀ value used for postoperative brain tumors was"0.75" and TCD ₅₀ as 27.04 Gy ^[15] was used for this study. For NTCP, appropriate model is required to analyze the effects extensively. This study limits up to evaluate injury in terms of BED acute toxicity of NSC and BED late effects of brain without compromising plan quality.

BED is expressed as

Based on study by Marsh et al., ^[16] BED were calculated by using mean dose "d" per fraction because the goal of current investigation is to use DVH as a relative measure of biological response for assessing acute effects of NSC and late effects of non-NSC parts of brain and brainstem. ^[17]

Results

(a) Physical indices

The extracted header files of group 1 plans are modulated by changing the priority values from -x to +x for NSC in the optimization schemes, which corresponds to the group 2 planning indices less than standard deviation (SD) 0.3%. The objective of the model is to figure out the output file, and it implies upon the equipoise between group 1 and group 2 plans to elucidate the potential sparing without compromising plan quality. In this action of the RTOG indices, the calculated value of QC _mean is 0.878 and 0.883, HI _mean is 1.088 and 1.074, CI _mean is 1.146 and 1.144, and CNmean is 0.613 and 0.611, as shown in [Figure - 4] and [Table - 1]. The acceptable SDs for various physical indices are shown in [Table - 2]. We found that the computation model is well within the limit in terms of equipoise criteria adopted for the model.

(b) Biological indices

[Table - 3] clearly shows a reduction in acute and late reaction of NSC acute effects and brain while achieving similar EUD and TCP in both groups. Mean EUD in group 1 is 51.47 and it is 51.60 in group 2. Mean variation is 0.10% for TCP and shows agreement for the control probability for the adopted model. The mean and maximum reduction of 12.23% and 28.33% respectively on stem cells and 14.69% and 33.56% respectively on late effects of brain were observed, as shown in [Table - 3]; BED-DVH Histogram of NSC and Target and isodose distribution of transverse view are shown in [Figure - 5] and [Figure - 6].

Discussion

The overall goal of this study is to incorporate patient-specific biological information and plan-specific indices on to multivariative analysis platform. The current model was applied in the stem cell sparing IMRT plans for Glioblastoma Multiforme (GBM). It has been speculated that the irradiation of NSC results in the inability to repair radiation-induced damage to normal brain tissue. The phenotypic expression of damage due to normal tissue may manifest as memory loss and loss of executive function. ^[11],[18] Pediatric patients undergoing brain irradiation may lead to hearing loss, severe global cognitive deficiencies, and neuroendocrine deficits. ^{[19],[20],[21],[22]} Involvement of NSC in this series varies from 18 to 87% in tumor volume, depending on the volume and site of the tumor, and OAR dose limits are maintained during the plans, as represented in [Figure - 2]. There is dearth of evidence about objective and quantitative correlation regarding neurocognitive decline in 5-year survival statistics. There is a lack of information regarding TD 50/5 for human NSC′s. Hence, it is difficult to generate accurate NTCP models for this population. ^[16] Predicting local control for the planned distribution is unique for each plan.

Our study demonstrates the feasibility of creating nonhomogenous distribution using IMRT techniques such that higher therapeutic ratio was achieved while aiming for higher EUD and TCP. The design of nonuniform dose distribution was aided by information gleaned from various physical and biological indices. This basic version of computational tool is not a stand-alone executable environment. It requires repeated DVH input files at each phase to run this tool. It establishes a useful platform for sparing of NSC with benefits of steep dose gradients using IMRT planning. Similar planning can also be carried out by this tool. Extended works of NSC includes more patient data points, and tools need to be modified with additional features to analyze patient response, neurocognitive tests, updated scoring system like National Cancer Institute Common Toxicity Criteria version, RTOG/ European Organization for Research and Treatment of Cancer (EORTC) protocols, ^[23] and multi-institutional clinical trial results.

Conclusions

Based on the evidence from in vitro animal and human studies which support the hypothesis of radiation-induced damage to NSC compartment, ^{[24],[25],[26],[27],[28],[29],[30],[31]} our results support the recent investigators ^[32] to emphasize the avoidance of NSC compartment. Its validation is recommended as an important parameter to be included in IMRT plans for partial-brain irradiation. The pilot study has shown that it is feasible to use RTOG-recommended indices and biological indices in planning on this platform. It is however necessary to validate this model as well on a larger cohort of patients and correlate it to subsequent clinical benefits.

References

Copyright 2011 - Journal of Cancer Research and Therapeutics

The following images related to this document are available:

Photo images