Reporting and analyzing dose distributions: a concept of equivalent uniform dose

Relationships between Regional Radiation Doses and Cognitive Decline in Children Treated with Cranio-Spinal Irradiation for Posterior Fossa Tumors

Pediatric posterior fossa tumor (PFT) survivors who have been treated with cranial radiation therapy often suffer from cognitive impairments that might relate to IQ decline. Radiotherapy (RT) distinctly affects brain regions involved in different cognitive functions. However, the relative contribution of regional irradiation to the different cognitive impairments still remains unclear. We investigated the relationships between the changes in different cognitive scores and radiation dose distribution in 30 children treated for a PFT. Our exploratory analysis was based on a principal component analysis (PCA) and an ordinary least square regression approach. The use of a PCA was an innovative way to cluster correlated irradiated regions due to similar radiation therapy protocols across patients. Our results suggest an association between working memory decline and a high dose (equivalent uniform dose, EUD) delivered to the orbitofrontal regions, whereas the decline of processing speed seemed more related to EUD in the temporal lobes and posterior fossa. To identify regional effects of RT on cognitive functions may help to propose a rehabilitation program adapted to the risk of cognitive impairment.

The use of TCP based EUD to rank and compare lung radiotherapy plans: in-silico study to evaluate the correlation between TCP with physical quality indices

BACKGROUND

To apply the equivalent uniform dose (EUD) radiobiological model to estimate the tumor control probability (TCP) scores for treatment plans using different radiobiological parameter settings, and to evaluate the correlation between TCP and physical quality indices of the treatment plans.

METHODS

Ten radiotherapy treatment plans for lung cancer were generated. The dose distributions were calculated using anisotropic analytical algorithm (AAA). Dose parameters and quality indices derived from dose volume histograms (DVH) for target volumes were evaluated. The predicted TCP was computed using EUD model with tissue-specific parameter (a=-10). The assumed radiobiological parameter setting for adjuvant therapy [tumor dose to control 50% of the tumor (TCD₅₀) =36.5 Gy and γ₅₀=0.72] and curative intent (TCD₅₀=51.24 Gy and γ₅₀=0.83) were used. The bootstrap method was used to estimate the 95% confidence interval (95% CI). The coefficients (ρ) from Spearman's rank test were calculated to assess the correlation between quality indices with TCP. Wilcoxon paired test was used to calculate P value.

RESULTS

The 95% CI of TCP were 70.6-81.5 and 46.6-64.7, respectively, for adjuvant radiotherapy and curative intent. The TCP outcome showed a positive and good correlation with calculated dose to 95% of the target volume (D95%) and minimum dose (Dmin). Consistently, TCP correlate negatively with heterogeneity indices.

CONCLUSIONS

This study confirms that more relevant and robust radiobiological parameters setting should be integrated according to cancer type. The positive correlation with quality indices gives chance to improve the clinical out-come by optimizing the treatment plans to maximize the Dmin and D95%. This attempt to increase the TCP should be carried out with the respect of dose constraints for organs at risks. However, the negative correlation with heterogeneity indices shows that the optimization of beam arrangements could be also useful. Attention should be paid to obtain an appropriate optimization of initial plans, when comparing and ranking radiotherapy plans using TCP models, to avoid over or underestimated for TCP outcome.

Modelling duodenum radiotherapy toxicity using cohort dose-volume-histogram data

Background and purpose

Gastro-intestinal toxicity is dose-limiting in abdominal radiotherapy and correlated with duodenum dose-volume parameters. We aimed to derive updated NTCP model parameters using published data and prospective radiotherapy quality-assured cohort data.

Material and methods

A systematic search identified publications providing duodenum dose-volume histogram (DVH) statistics for clinical studies of conventionally-fractionated radiotherapy. Values for the Lyman-Kutcher-Burman (LKB) NTCP model were derived through sum-squared-error minimisation and using leave-one-out cross-validation. Data were corrected for fraction size and weighted according to patient numbers, and the model refined using individual patient DVH data for two further cohorts from prospective clinical trials.

Results

Six studies with published DVH data were utilised, and with individual patient data included outcomes for 531 patients in total (median follow-up 16 months). Observed gastro-intestinal toxicity rates ranged from 0% to 14% (median 8%). LKB parameter values for unconstrained fit to published data were: n = 0.070, m = 0.46, TD₅₀₍₁₎ [Gy] = 183.8, while the values for the model incorporating the individual patient data were n = 0.193, m = 0.51, TD₅₀₍₁₎ [Gy] = 299.1.

Conclusions

LKB parameters derived using published data are shown to be consistent to those previously obtained using individual patient data, supporting a small volume-effect and dependence on exposure to high threshold dose.

Measurement and analysis of the impact of time-interval, temperature and radiation dose on tumour cell survival and its application in thermoradiotherapy plan evaluation

PURPOSE

Biological modelling of thermoradiotherapy may further improve patient selection and treatment plan optimisation, but requires a model that describes the biological effect as a function of variables that affect treatment outcome (e.g. temperature, radiation dose). This study aimed to establish such a model and its parameters. Additionally, a clinical example was presented to illustrate the application.

METHODS

Cell survival assays were performed at various combinations of radiation dose (0-8 Gy), temperature (37-42 °C), time interval (0-4 h) and treatment sequence (radiotherapy before/after hyperthermia) for two cervical cancer cell lines (SiHa and HeLa). An extended linear-quadratic model was fitted to the data using maximum likelihood estimation. As an example application, a thermoradiotherapy plan (23 × 2 Gy + weekly hyperthermia) was compared with a radiotherapy-only plan (23 × 2 Gy) for a cervical cancer patient. The equivalent uniform radiation dose (EUD) in the tumour, including confidence intervals, was estimated using the SiHa parameters. Additionally, the difference in tumour control probability (TCP) was estimated.

RESULTS

Our model described the dependency of cell survival on dose, temperature and time interval well for both SiHa and HeLa data (R²=0.90 and R²=0.91, respectively), making it suitable for biological modelling. In the patient example, the thermoradiotherapy plan showed an increase in EUD of 9.8 Gy that was robust (95% CI: 7.7-14.3 Gy) against propagation of the uncertainty in radiobiological parameters. This corresponded to a 20% (95% CI: 15-29%) increase in TCP.

CONCLUSIONS

This study presents a model that describes the cell survival as a function of radiation dose, temperature and time interval, which is essential for biological modelling of thermoradiotherapy treatments.

Fast and fuzzy multi-objective radiotherapy treatment plan generation for head and neck cancer patients with the lexicographic reference point method (LRPM)

Previously, we have proposed Erasmus-iCycle, an algorithm for fully automated IMRT plan generation based on prioritised (lexicographic) multi-objective optimisation with the 2-phase ϵ-constraint (2pϵc) method. For each patient, the output of Erasmus-iCycle is a clinically favourable, Pareto optimal plan. The 2pϵc method uses a list of objective functions that are consecutively optimised, following a strict, user-defined prioritisation. The novel lexicographic reference point method (LRPM) is capable of solving multi-objective problems in a single optimisation, using a fuzzy prioritisation of the objectives. Trade-offs are made globally, aiming for large favourable gains for lower prioritised objectives at the cost of only slight degradations for higher prioritised objectives, or vice versa. In this study, the LRPM is validated for 15 head and neck cancer patients receiving bilateral neck irradiation. The generated plans using the LRPM are compared with the plans resulting from the 2pϵc method. Both methods were capable of automatically generating clinically relevant treatment plans for all patients. For some patients, the LRPM allowed large favourable gains in some treatment plan objectives at the cost of only small degradations for the others. Moreover, because of the applied single optimisation instead of multiple optimisations, the LRPM reduced the average computation time from 209.2 to 9.5 min, a speed-up factor of 22 relative to the 2pϵc method.

Delivered dose can be a better predictor of rectal toxicity than planned dose in prostate radiotherapy

Background and purpose

For the first time, delivered dose to the rectum has been calculated and accumulated throughout the course of prostate radiotherapy using megavoltage computed tomography (MVCT) image guidance scans. Dosimetric parameters were linked with toxicity to test the hypothesis that delivered dose is a stronger predictor of toxicity than planned dose.

Material and methods

Dose–surface maps (DSMs) of the rectal wall were automatically generated from daily MVCT scans for 109 patients within the VoxTox research programme. Accumulated-DSMs, representing total delivered dose, and planned-DSMs, from planning CT data, were parametrised using Equivalent Uniform Dose (EUD) and ‘DSM dose-width’, the lateral dimension of an ellipse fitted to a discrete isodose cluster. Associations with 6 toxicity endpoints were assessed using receiver operator characteristic curve analysis.

Results

For rectal bleeding, the area under the curve (AUC) was greater for accumulated dose than planned dose for DSM dose-widths up to 70 Gy. Accumulated 65 Gy DSM dose-width produced the strongest spatial correlation (AUC 0.664), while accumulated EUD generated the largest AUC overall (0.682). For proctitis, accumulated EUD was the only reportable predictor (AUC 0.673). Accumulated EUD was systematically lower than planned EUD.

Conclusions

Dosimetric parameters extracted from accumulated DSMs have demonstrated stronger correlations with rectal bleeding and proctitis, than planned DSMs.

Comparison between the four-field box and field-in-field techniques for conformal radiotherapy of the esophagus using dose-volume histograms and normal tissue complication probabilities

PURPOSE

We evaluated and compared the performance of the field-in-field (FIF) to that of the four-field box (4FB) technique regarding dosimetric and radiobiological parameters for radiotherapy of esophageal carcinoma.

MATERIALS AND METHODS

Twenty patients with esophageal cancer were selected. For each patient, two treatment plans were created: 4FB and FIF. The parameters compared included the conformity index (CI), homogeneity index (HI), D _mean, D _max, tumor control probability (TCP), V _20Gy and V _30Gy of the heart and lungs, normal tissue complication probability (NTCP), and monitor units per fraction (MU/fr).

RESULTS

A paired t-test analysis did not show any significant differences (p > 0.05) between the two techniques in terms of the CI and TCP. However, the HI significantly improved when the FIF was applied. D _max of the PTV, lung, and spinal cord were also significantly better with the FIF. Moreover, the lung V _20Gy as well as the NTCPs of the lung and spinal cord significantly reduced when the FIF was used, and the MU/fr was significantly decreased.

CONCLUSIONS

The FIF showed evident advantages over 4FB: a more homogeneous dose distribution, lower D _max values, and fewer required MUs, while it also retained PTV dose conformality. FIF should be considered as a simple technique to use clinically in cases with esophageal malignancies, especially in clinics with no IMRT.

Theoretical effectiveness of cell survival in fractionated radiotherapy with hypoxia-targeted dose escalation

PURPOSE

The goal of this article is to compute the cell survival during fractionated radiotherapy with non-uniform hypoxia-targeted dose distribution relative to cell survival for a uniform dose distribution with equal integral tumor dose. The analysis is performed for different parameters of radiotherapy with conventional and hypofractionated dose regimens.

METHODS

Our analysis is done using a parsimonious tumor response model that describes the major components of tumor response to radiotherapy such as radiosensitivity, cell proliferation, and hypoxia using as few variables as possible. Two levels of oxygenated and hypoxic cells with the survival curves described by the linear quadratic (LQ) model are implemented in the model. The model allows for analytical solutions for relative cell survival in a single fraction simulation which can be used for additional validation of our numerical simulations. The relative cell survival was computed for conventional and hypofractionated dose regimens in a model problem with radiobiological parameters for the non-small-cell lung cancer. Sensitivity of cell survival to different parameters of radiotherapy such as the relative volume of hypoxic fraction, boost dose ratio, re-oxygenation rate, and delivery errors was investigated.

RESULTS

Our computational and analytical results show that relative cell survival for non-uniform and uniform dose distributions is larger than 1.0 during the first few fractions of radiotherapy with conventional fractionation. This indicates that the uniform dose distribution produces a higher cell killing effect for the equal integral dose. This may stem from domination of linear contribution to the cell killing effect seen in the dose range of 1-2 Gy and a steeper slope of survival curve in the oxygenated tumor region. This effect can only happen if the distribution of clonogens is nearly uniform; therefore, after the first few fractions, the non-uniform dose distributions outperform the uniform dose distribution and relative cell survival becomes less than 1.0. However, re-oxygenation and delivery errors can also reduce the effectiveness of hypoxia-targeted non-uniform dose distributions toward the end of treatment. For hypofractionated radiotherapy with fractional dose >3 Gy, the relative cell survival was always below 1.0, which means the non-uniform dose distributions produced higher cell killing effect than the uniform dose distribution during the entire treatment.

CONCLUSION

The data obtained in this work suggest that non-uniform hypoxia-targeted dose distributions are less effective at cell killing than uniform dose distributions at the beginning of radiotherapy with conventional fractionation. However; non-uniform dose distributions can outperform uniform dose distribution by the end of the treatment if the effects of re-oxygenation and delivery errors are reduced. In hypofractionated radiotherapy, non-uniform hypoxia-targeted dose distributions are more efficient than uniform dose distributions during the entire treatment.

Risk of secondary cancers: Bridging epidemiology and modeling

Epidemiological studies of long term radiotherapy survivors provide useful insights into dose-response relationships for secondary cancer induction risk at high doses. There are uncertainties involved in estimating the dose to the location of the second malignancy, because the dose distributions in radiotherapy patients can be spatially highly heterogeneous and the size of the diagnosed tumor is on the order of a few cm. Therefor it is nearly impossible to obtain the exact dose corresponding to the exact tumor induction location and so organ specific dose-response relationships have large errors not only in the reported risk, but also in the estimated dose. In this work two alternative methods are proposed for future applications involving investigations into dose response relationships for second cancer induction risk, the method of organ sub-division and the method of risk equivalent dose. The method of organ sub-division takes the inevitable inhomogeneous dose distribution into account by applying epidemiological methods to organ sub-divisions which have a nearly homogenous dose. The method of risk equivalent dose combines risk modeling and epidemiological data analysis. Risk models can be optimized by using an iterative procedure assuming a variation of organ specific dose-responses. The advantage of the alternative methods is that the inhomogeneity of the dose in the organs at risk is taken into account. The second method has the additional advantage that the dose to the location of the tumor site must not be known and that epidemiologically obtained risks that were not stratified by organ specific risk can be used.

Toward a better prescription method for external radiotherapy

Aim: The aim of the study was to compare several methods of dose prescription, the mean dose, the median dose, the effective dose and the generalized Equivalent Uniform Dose (gEUD).

Background: The dose distribution in the planning target volume is never fully homogenous. Depending on the dose prescription method for the same prescribed dose different biologically equivalent doses are delivered. The latest ICRU Report 83 proposes to prescribe the dose to the median dose in the PTV. Several other methods are also in common use. It is important to know what are differences of doses actually delivered depending on the dose prescription method.

Materials and methods: The study was performed for three groups of patients treated radically with external beams in Brzozow, over the 2012-2013 period. The groups were of patients with breast, lung and prostate cancer. There were 10 patients in each group. For each patient all metrics, i.e. the mean dose, the median dose, the effective dose and the generalized Equivalent Uniform Dose, were calculated. The influence of the dose homogeneity in the PTV on the results is also evaluated. The gEUD was used as a reference dose prescription method.

Results: For all patients, an almost perfect correlation between the median dose and the gEUD was obtained. Worse correlation was obtained between other metrics and the gEUD. The median dose is almost always a little higher than the gEUD, but the ratio of these two values never exceeded 1.013.

Conclusion: The median dose seems to be a good and simple method of dose prescription.

Risks and benefits of reducing target volume margins in breast tangent radiotherapy

This study investigates the potential benefits of planning target volume (PTV) margin reduction for whole breast radiotherapy in relation to dose received by organs at risk (OARs), as well as reductions in radiation-induced secondary cancer risk. Such benefits were compared to the increased radiation-induced secondary cancer risk attributed from increased ionizing radiation imaging doses. Ten retrospective patients' computed tomography datasets were considered. Three computerized treatment plans with varied PTV margins (0, 5 and 10 mm) were created for each patient complying with the Radiation Therapy Oncology Group (RTOG) 1005 protocol requirements. The BEIR VII lifetime attributable risk (LAR) model was used to estimate secondary cancer risk to OARs. The LAR was assessed for all treatment plans considering (a) doses from PTV margin variation and (b) doses from two (daily and weekly) kilovoltage cone beam computed tomography (kV CBCT) imaging protocols during the course of treatment. We found PTV margins from largest to smallest resulted in a mean OAR relative dose reduction of 31% (heart), 28% (lung) and 23% (contralateral breast) and the risk of radiation-induced secondary cancer by a relative 23% (contralateral breast) and 22% (contralateral lung). Daily image-guidance using kV CBCT increased the risk of radiation induced secondary cancer to the contralateral breast and contralateral lung by a relative 1.6-1.9% and 1.9-2.5% respectively. Despite the additional dose from kV CBCT for the two considered imaging protocols, smaller PTV margins would still result in an overall reduction in secondary cancer risk.

Statistical-learning strategies generate only modestly performing predictive models for urinary symptoms following external beam radiotherapy of the prostate: A comparison of conventional and machine-learning methods

PURPOSE

Given the paucity of available data concerning radiotherapy-induced urinary toxicity, it is important to ensure derivation of the most robust models with superior predictive performance. This work explores multiple statistical-learning strategies for prediction of urinary symptoms following external beam radiotherapy of the prostate.

METHODS

The performance of logistic regression, elastic-net, support-vector machine, random forest, neural network, and multivariate adaptive regression splines (MARS) to predict urinary symptoms was analyzed using data from 754 participants accrued by TROG03.04-RADAR. Predictive features included dose-surface data, comorbidities, and medication-intake. Four symptoms were analyzed: dysuria, haematuria, incontinence, and frequency, each with three definitions (grade ≥ 1, grade ≥ 2 and longitudinal) with event rate between 2.3% and 76.1%. Repeated cross-validations producing matched models were implemented. A synthetic minority oversampling technique was utilized in endpoints with rare events. Parameter optimization was performed on the training data. Area under the receiver operating characteristic curve (AUROC) was used to compare performance using sample size to detect differences of ≥0.05 at the 95% confidence level.

RESULTS

Logistic regression, elastic-net, random forest, MARS, and support-vector machine were the highest-performing statistical-learning strategies in 3, 3, 3, 2, and 1 endpoints, respectively. Logistic regression, MARS, elastic-net, random forest, neural network, and support-vector machine were the best, or were not significantly worse than the best, in 7, 7, 5, 5, 3, and 1 endpoints. The best-performing statistical model was for dysuria grade ≥ 1 with AUROC ± standard deviation of 0.649 ± 0.074 using MARS. For longitudinal frequency and dysuria grade ≥ 1, all strategies produced AUROC>0.6 while all haematuria endpoints and longitudinal incontinence models produced AUROC<0.6.

CONCLUSIONS

Logistic regression and MARS were most likely to be the best-performing strategy for the prediction of urinary symptoms with elastic-net and random forest producing competitive results. The predictive power of the models was modest and endpoint-dependent. New features, including spatial dose maps, may be necessary to achieve better models.

The Impact of the Grid Size on TomoTherapy for Prostate Cancer

Discretization errors due to the digitization of computed tomography images and the calculation grid are a significant issue in radiation therapy. Such errors have been quantitatively reported for a fixed multifield intensity-modulated radiation therapy using traditional linear accelerators. The aim of this study is to quantify the influence of the calculation grid size on the dose distribution in TomoTherapy. This study used ten treatment plans for prostate cancer. The final dose calculation was performed with "fine" (2.73 mm) and "normal" (5.46 mm) grid sizes. The dose distributions were compared from different points of view: the dose-volume histogram (DVH) parameters for planning target volume (PTV) and organ at risk (OAR), the various indices, and dose differences. The DVH parameters were used Dmax, D2%, D2cc, Dmean, D95%, D98%, and Dmin for PTV and Dmax, D2%, and D2cc for OARs. The various indices used were homogeneity index and equivalent uniform dose for plan evaluation. Almost all of DVH parameters for the "fine" calculations tended to be higher than those for the "normal" calculations. The largest difference of DVH parameters for PTV was Dmax and that for OARs was rectal D2cc. The mean difference of Dmax was 3.5%, and the rectal D2cc was increased up to 6% at the maximum and 2.9% on average. The mean difference of D95% for PTV was the smallest among the differences of the other DVH parameters. For each index, whether there was a significant difference between the two grid sizes was determined through a paired t-test. There were significant differences for most of the indices. The dose difference between the "fine" and "normal" calculations was evaluated. Some points around high-dose regions had differences exceeding 5% of the prescription dose. The influence of the calculation grid size in TomoTherapy is smaller than traditional linear accelerators. However, there was a significant difference. We recommend calculating the final dose using the "fine" grid size.

Grid Block Design Based on Monte Carlo Simulated Dosimetry, the Linear Quadratic and Hug-Kellerer Radiobiological Models

Purpose

The clinical efficacy of Grid therapy has been examined by several investigators. In this project, the hole diameter and hole spacing in Grid blocks were examined to determine the optimum parameters that give a therapeutic advantage.

Methods

The evaluations were performed using Monte Carlo (MC) simulation and commonly used radiobiological models. The Geant4 MC code was used to simulate the dose distributions for 25 different Grid blocks with different hole diameters and center-to-center spacing. The therapeutic parameters of these blocks, namely, the therapeutic ratio (TR) and geometrical sparing factor (GSF) were calculated using two different radiobiological models, including the linear quadratic and Hug-Kellerer models. In addition, the ratio of the open to blocked area (ROTBA) is also used as a geometrical parameter for each block design. Comparisons of the TR, GSF, and ROTBA for all of the blocks were used to derive the parameters for an optimum Grid block with the maximum TR, minimum GSF, and optimal ROTBA. A sample of the optimum Grid block was fabricated at our institution. Dosimetric characteristics of this Grid block were measured using an ionization chamber in water phantom, Gafchromic film, and thermoluminescent dosimeters in Solid Water™ phantom materials.

Results

The results of these investigations indicated that Grid blocks with hole diameters between 1.00 and 1.25 cm and spacing of 1.7 or 1.8 cm have optimal therapeutic parameters (TR > 1.3 and GSF~0.90). The measured dosimetric characteristics of the optimum Grid blocks including dose profiles, percentage depth dose, dose output factor (cGy/MU), and valley-to-peak ratio were in good agreement (±5%) with the simulated data.

Conclusion

In summary, using MC-based dosimetry, two radiobiological models, and previously published clinical data, we have introduced a method to design a Grid block with optimum therapeutic response. The simulated data were reproduced by experimental data.

Dosimetric and radiobiological comparison of Cyberknife and Tomotherapy in stereotactic body radiotherapy for localized prostate cancer

BACKGROUND AND PURPOSE

As recent studies have suggested relatively low α/β for prostate cancer, the interest in hypofractionated stereotactic body radiotherapy (SBRT) for prostate cancer is rising. The aim of this study is to compare dosimetric results of Cyberknife (CK) with Tomotherapy (HT) in SBRT for localized prostate cancer. Furthermore, the radiobiologic consequences of heterogeneous dose distribution are also analyzed.

MATERIAL AND METHOD

A total of 12 cases of localized prostate cancer previously treated with SBRT were collected. Treatments had been planned and delivered using CK. Then HT plans were generated for comparison afterwards. The prescribed dose was 37.5Gy in 5 fractions. Dosimetric indices for target volumes and organs at risk (OAR) were compared. For radiobiological evaluation, generalized equivalent uniform dose (gEUD) and normal tissue complication probability (NTCP) were calculated and compared.

RESULT

Both CK and HT achieved target coverage while meeting OAR constraints adequately. HT plans resulted in better dose homogeneity (Homogeneity index: 1.04±0.01 vs. 1.21±0.01; p = 0.0022), target coverage (97.74±0.86% vs. 96.56±1.17%; p = 0.0076) and conformity (new vonformity index: 1.16±0.05 vs. 1.21±0.04; p = 0.0096). HT was shown to predict lower late rectal toxicity as compared to CK. Integral dose to body was also significantly lower in HT plans (46.59±6.44 Gy'L vs 57.05±11.68 Gy'L; p = 0.0029).

CONCLUSION

Based on physical dosimetry and radiobiologic considerations, HT may have advantages over CK, specifically in rectal sparing which could translate into clinical benefit of decreased late toxicities.

Impact of dose calculation models on radiotherapy outcomes and quality adjusted life years for lung cancer treatment: do we need to measure radiotherapy outcomes to tune the radiobiological parameters of a normal tissue complication probability model?

BACKGROUND

The equivalent uniform dose (EUD) radiobiological model can be applied for lung cancer treatment plans to estimate the tumor control probability (TCP) and the normal tissue complication probability (NTCP) using different dose calculation models. Then, based on the different calculated doses, the quality adjusted life years (QALY) score can be assessed versus the uncomplicated tumor control probability (UTCP) concept in order to predict the overall outcome of the different treatment plans.

METHODS

Nine lung cancer cases were included in this study. For the each patient, two treatments plans were generated. The doses were calculated respectively from pencil beam model, as pencil beam convolution (PBC) turning on 1D density correction with Modified Batho's (MB) method, and point kernel model as anisotropic analytical algorithm (AAA) using exactly the same prescribed dose, normalized to 100% at isocentre point inside the target and beam arrangements. The radiotherapy outcomes and QALY were compared. The bootstrap method was used to improve the 95% confidence intervals (95% CI) estimation. Wilcoxon paired test was used to calculate P value.

RESULTS

Compared to AAA considered as more realistic, the PBC_MB overestimated the TCP while underestimating NTCP, P<0.05. Thus the UTCP and the QALY score were also overestimated.

CONCLUSIONS

To correlate measured QALY's obtained from the follow-up of the patients with calculated QALY from DVH metrics, the more accurate dose calculation models should be first integrated in clinical use. Second, clinically measured outcomes are necessary to tune the parameters of the NTCP model used to link the treatment outcome with the QALY. Only after these two steps, the comparison and the ranking of different radiotherapy plans would be possible, avoiding over/under estimation of QALY and any other clinic-biological estimates.

Correlation between pneumonitis risk in radiation oncology and lung density measured with X-ray computed tomography

BACKGROUND

The risk of toxicity with radiation oncology for lung cancer limits the maximal radiation dose that can be delivered to thoracic tumors. This study aims at investigating the correlation between normal tissue complication probability (NTCP) and physical lung density by analyzing the computed tomography (CT) scan imaging used for radiotherapy dose planning.

METHODS

Data from CT of lung cancer patients (n=10), treated with three dimensional radiotherapy, were selected for this study. The dose was calculated using analytical anisotropic algorithm (AAA). Dose volume histograms (DVH) for healthy lung (lung excluding targets) were calculated. The NTCP for lung radiation induced pneumonitis was computed using initial radiobiological parameters from Lyman-Kutcher and Burman (LKB) model and readjusted parameters for AAA, with α/β=3. The correlation coefficient "rho" was calculated using Spearman's rank test. The bootstrap method was used to estimate the 95% confidence interval (95% CI). Wilcoxon paired test was used to calculate P values.

RESULTS

Bootstrapping simulation revealed significant difference between NTCP computed with the initial radiobiological parameters and that computed with the parameters readjusted for AAA (P=0.03). The results of simulations based on 1,000 replications showed no correlation for NTCP with density, with "rho" <0.3.

CONCLUSIONS

For a given set of patients, we assessed the correlation between NTCP and lung density using bootstrap analysis. The lack of correlation could result either from a very accurate dose calculation, by AAA, whatever the lung density yielding a NTCP result only dependant of the dose and not any more of the density; or to the very limited range of natural variation of relative electronic density (0.15 to 0.20) observed in this small series of patients. Another important parameter is the bootstrap simulation with 1,000 random samplings may have underestimated the correlation, since the initial data (n=10) showed a weak correlation.

Dosimetric and radiobiological comparison of TG-43 and Monte Carlo calculations in 192Ir breast brachytherapy applications

PURPOSE

To investigate the clinical significance of introducing model based dose calculation algorithms (MBDCAs) as an alternative to TG-43 in ¹⁹²Ir interstitial breast brachytherapy.

MATERIALS AND METHODS

A 57 patient cohort was used in a retrospective comparison between TG-43 based dosimetry data exported from a treatment planning system and Monte Carlo (MC) dosimetry performed using MCNP v. 6.1 with plan and anatomy information in DICOM-RT format. Comparison was performed for the target, ipsilateral lung, heart, skin, breast and ribs, using dose distributions, dose-volume histograms (DVH) and plan quality indices clinically used for plan evaluation, as well as radiobiological parameters.

RESULTS

TG-43 overestimation of target DVH parameters is statistically significant but small (less than 2% for the target coverage indices and 4% for homogeneity indices, on average). Significant dose differences (>5%) were observed close to the skin and at relatively large distances from the implant leading to a TG-43 dose overestimation for the organs at risk. These differences correspond to low dose regions (<50% of the prescribed dose), being less than 2% of the prescribed dose. Detected dosimetric differences did not induce clinically significant differences in calculated tumor control probabilities (mean absolute difference <0.2%) and normal tissue complication probabilities.

CONCLUSION

While TG-43 shows a statistically significant overestimation of most indices used for plan evaluation, differences are small and therefore not clinically significant. Improved MBDCA dosimetry could be important for re-irradiation, technique inter-comparison and/or the assessment of secondary cancer induction risk, where accurate dosimetry in the whole patient anatomy is of the essence.

New approach in lung cancer radiotherapy offers better normal tissue sparing

PURPOSE

Medical images are more than pictures. They contain additional quantitative information which can be interrogated, quantified, and utilized. Besides anatomical information computed tomography (CT) imaging data provide electron density information. Radiotherapy use of this density information is limited to its application only in dose calculations. The direct product of dose, density, and volume forms a quantity called integral dose. The integral dose delivered to a volume of interest is the total energy deposited in that volume. Here it is hypothesized that minimization of the integral dose is advantageous in radiotherapy planning. The purpose of this work is to study the incorporation of quantitative imaging information in radiotherapy inverse optimization through total energy minimization (Energy hereafter).

DESIGN

Twenty lung patient plans were studied. For each patient density was quantified on voxel-by-voxel basis through image gray value-to-density conversion curves. Energy-based objective function was used for inverse radiotherapy plan optimization. The obtained plans were evaluated in the light of current standard of care, based on dose-volume (DVH) optimization approach.

RESULTS

The statistical significance analyses of the results indicated that the doses to normal tissue were between 14% and 45% lower, when Energy-based optimization was used instead of DVH-based optimization.

CONCLUSION

Incorporation of quantitative imaging information, through CT derived density, in the optimization cost function allows reduction of dose to normal tissue for NSCLC cases. Energy-based radiotherapy plans result in lower normal tissue dose and potentially lower complication rates compared to standard of care.

Comparison of treatment plans: a retrospective study by the method of radiobiological evaluation

There are many situations in radiotherapy where multiple treatment plans need to be compared for selection of an optimal plan. In this study we performed the radiobiological method of plan evaluation to verify the treatment plan comparison procedure of our clinical practice. We estimated and correlated various radiobiological dose indices with physical dose metrics for a total of 30 patients representing typical cases of head and neck, prostate and brain tumors. Three sets of plans along with a clinically approved plan (final plan) treated by either Intensity Modulated Radiation Therapy (IMRT) or Rapid Arc (RA) techniques were considered. The study yielded improved target coverage for final plans, however, no appreciable differences in doses and the complication probabilities of organs at risk were noticed. Even though all four plans showed adequate dose distributions, from dosimetric point of view, the final plan had more acceptable dose distribution. The estimated biological outcome and dose volume histogram data showed least differences between plans for IMRT when compared to RA. Our retrospective study based on 120 plans, validated the radiobiological method of plan evaluation. The tumor cure or normal tissue complication probabilities were found to be correlated with the corresponding physical dose indices.

Dosimetric and radiobiological characterizations of prostate intensity-modulated radiotherapy and volumetric-modulated arc therapy: A single-institution review of ninety cases

This study reviewed prostate volumetric-modulated arc therapy (VMAT) plans with intensity-modulated radiotherapy (IMRT) plans after prostate IMRT technique was replaced by VMAT in an institution. Characterizations of dosimetry and radiobiological variation in prostate were determined based on treatment plans of 40 prostate IMRT patients (planning target volume = 77.8–335 cm³) and 50 VMAT patients (planning target volume = 120–351 cm³) treated before and after 2013, respectively. Both IMRT and VMAT plans used the same dose-volume criteria in the inverse planning optimization. Dose-volume histogram, mean doses of target and normal tissues (rectum, bladder and femoral heads), dose-volume points (D_99% of planning target volume; D_30%, D_50%, V_{30 Gy} and V_{35 Gy} of rectum and bladder; D_5%, V_{14 Gy}, V_{22 Gy} of femoral heads), conformity index (CI), homogeneity index (HI), gradient index (GI), prostate tumor control probability (TCP), and rectal normal tissue complication probability (NTCP) based on the Lyman-Burman-Kutcher algorithm were calculated for each IMRT and VMAT plan. From our results, VMAT plan was found better due to its higher (1.05%) CI, lower (0.83%) HI and (0.75%) GI than IMRT. Comparing doses in normal tissues between IMRT and VMAT, it was found that IMRT mostly delivered higher doses of about 1.05% to the normal tissues than VMAT. Prostate TCP and rectal NTCP were found increased (1%) for VMAT than IMRT. It is seen that VMAT technique can decrease the dose-volume evaluation criteria for the normal tissues. Based on our dosimetric and radiobiological results in treatment plans, it is concluded that our VMAT implementation could produce comparable or slightly better target coverage and normal tissue sparing with a faster treatment time in prostate radiotherapy.

Multivariable normal tissue complication probability model-based treatment plan optimization for grade 2-4 dysphagia and tube feeding dependence in head and neck radiotherapy

BACKGROUND AND PURPOSE

Radiotherapy of the head and neck is challenged by the relatively large number of organs-at-risk close to the tumor. Biologically-oriented objective functions (OF) could optimally distribute the dose among the organs-at-risk. We aimed to explore OFs based on multivariable normal tissue complication probability (NTCP) models for grade 2-4 dysphagia (DYS) and tube feeding dependence (TFD).

MATERIALS AND METHODS

One hundred head and neck cancer patients were studied. Additional to the clinical plan, two more plans (an OF_DYS and OF_TFD-plan) were optimized per patient. The NTCP models included up to four dose-volume parameters and other non-dosimetric factors. A fully automatic plan optimization framework was used to optimize the OF_NTCP-based plans.

RESULTS

All OF_NTCP-based plans were reviewed and classified as clinically acceptable. On average, the Δdose and ΔNTCP were small comparing the OF_DYS-plan, OF_TFD-plan, and clinical plan. For 5% of patients NTCP_TFD reduced >5% using OF_TFD-based planning compared to the OF_DYS-plans.

CONCLUSIONS

Plan optimization using NTCP_DYS- and NTCP_TFD-based objective functions resulted in clinically acceptable plans. For patients with considerable risk factors of TFD, the OF_TFD steered the optimizer to dose distributions which directly led to slightly lower predicted NTCP_TFD values as compared to the other studied plans.

Strengths and Weaknesses of a Planar Whole-Body Method of (153)Sm Dosimetry for Patients with Metastatic Osteosarcoma and Comparison with Three-Dimensional Dosimetry

PURPOSE

Dosimetric accuracy depends directly upon the accuracy of the activity measurements in tumors and organs. The authors present the methods and results of a retrospective tumor dosimetry analysis in 14 patients with a total of 28 tumors treated with high activities of (153)Sm-ethylenediaminetetramethylenephosphonate ((153)Sm-EDTMP) for therapy of metastatic osteosarcoma using planar images and compare the results with three-dimensional dosimetry.

MATERIALS AND METHODS

Analysis of phantom data provided a complete set of parameters for dosimetric calculations, including buildup factor, attenuation coefficient, and camera dead-time compensation. The latter was obtained using a previously developed methodology that accounts for the relative motion of the camera and patient during whole-body (WB) imaging. Tumor activity values calculated from the anterior and posterior views of WB planar images of patients treated with (153)Sm-EDTMP for pediatric osteosarcoma were compared with the geometric mean value. The mean activities were integrated over time and tumor-absorbed doses were calculated using the software package OLINDA/EXM.

RESULTS

The authors found that it was necessary to employ the dead-time correction algorithm to prevent measured tumor activity half-lives from often exceeding the physical decay half-life of (153)Sm. Measured half-lives so long are unquestionably in error. Tumor-absorbed doses varied between 0.0022 and 0.27 cGy/MBq with an average of 0.065 cGy/MBq; however, a comparison with absorbed dose values derived from a three-dimensional analysis for the same tumors showed no correlation; moreover, the ratio of three-dimensional absorbed dose value to planar absorbed dose value was 2.19. From the anterior and posterior activity comparisons, the order of clinical uncertainty for activity and dose calculations from WB planar images, with the present methodology, is hypothesized to be about 70%.

CONCLUSION

The dosimetric results from clinical patient data indicate that absolute planar dosimetry is unreliable and dosimetry using three-dimensional imaging is preferable, particularly for tumors, except perhaps for the most sophisticated planar methods. The relative activity and patient kinetics derived from planar imaging show a greater level of reliability than the dosimetry.

Comparison of composite prostate radiotherapy plan doses with dependent and independent boost phases

Prostate cases commonly consist of dual phase planning with a primary plan followed by a boost. Traditionally, the boost phase is planned independently from the primary plan with the risk of generating hot or cold spots in the composite plan. Alternatively, boost phase can be planned taking into account the primary dose. The aim of this study was to compare the composite plans from independently and dependently planned boosts using dosimetric and radiobiological metrics. Ten consecutive prostate patients previously treated at our institution were used to conduct this study on the Raystation™ 4.0 treatment planning system. For each patient, two composite plans were developed: a primary plan with an independently planned boost and a primary plan with a dependently planned boost phase. The primary plan was prescribed to 54 Gy in 30 fractions to the primary planning target volume (PTV1) which includes prostate and seminal vesicles, while the boost phases were prescribed to 24 Gy in 12 fractions to the boost planning target volume (PTV2) that targets only the prostate. PTV coverage, max dose, median dose, target conformity, dose homogeneity, dose to OARs, and probabilities of benefit, injury, and complication-free tumor control (P+) were compared. Statistical significance was tested using either a 2-tailed Student's t-test or Wilcoxon signed-rank test. Dosimetrically, the composite plan with dependent boost phase exhibited smaller hotspots, lower maximum dose to the target without any significant change to normal tissue dose. Radiobiologically, for all but one patient, the percent difference in the P+ values between the two methods was not significant. A large percent difference in P+ value could be attributed to an inferior primary plan. The benefits of considering the dose in primary plan while planning the boost is not significant unless a poor primary plan was achieved.

SPIDERplan: A tool to support decision-making in radiation therapy treatment plan assessment

AIM

In this work, a graphical method for radiotherapy treatment plan assessment and comparison, named SPIDERplan, is proposed. It aims to support plan approval allowing independent and consistent comparisons of different treatment techniques, algorithms or treatment planning systems.

BACKGROUND

Optimized plans from modern radiotherapy are not easy to evaluate and compare because of their inherent multicriterial nature. The clinical decision on the best treatment plan is mostly based on subjective options.

MATERIALS AND METHODS

SPIDERplan combines a graphical analysis with a scoring index. Customized radar plots based on the categorization of structures into groups and on the determination of individual structures scores are generated. To each group and structure, an angular amplitude is assigned expressing the clinical importance defined by the radiation oncologist. Completing the graphical evaluation, a global plan score, based on the structures score and their clinical weights, is determined. After a necessary clinical validation of the group weights, SPIDERplan efficacy, to compare and rank different plans, was tested through a planning exercise where plans had been generated for a nasal cavity case using different treatment planning systems.

RESULTS

SPIDERplan method was applied to the dose metrics achieved by the nasal cavity test plans. The generated diagrams and scores successfully ranked the plans according to the prescribed dose objectives and constraints and the radiation oncologist priorities, after a necessary clinical validation process.

CONCLUSIONS

SPIDERplan enables a fast and consistent evaluation of plan quality considering all targets and organs at risk.

Assessing the shift of radiobiological metrics in lung radiotherapy plans using 2D gamma index

BACKGROUND

The purpose of this work is to investigate the 2D gamma (γ) maps to illustrate the change of radiobiological outcomes for lung radiotherapy plans and evaluate the correlation between tumor control probability (TCP), normal tissue complication probability (NTCP) with γ passing rates (γ-rates).

METHODS

Nine patients with lung cancer were used. The doses were calculated using Modified Batho method integrated with pencil beam convolution (MB-PBC) and anisotropic analytical algorithm (AAA) using the same beam arrangements and prescription dose. The TCP and NTCP were estimated, respectively, using equivalent uniform dose (EUD) model and Lyman-Kutcher-Burman (LKB) model. The correlation between ΔTCP or ΔNTCP with γ-rates, from 2%/2 and 3%/3 mm, were tested to explore the best correlation predicting the relevant γ criteria using Spearman's rank test (ρ). Wilcoxon paired test was used to calculate P value.

RESULTS

TCP value was significantly lower in the recalculated AAA plans as compared to MB plans. However, AAA predicted more NTCP on lung pneumonitis according to the LKB model and using relevant radiobiological parameters (n, m and TD50) for MB-PBC and AAA, with P=0.03. The data showed a weak correlation between radiobiological metrics with γ-rates or γ-mean, ρ<0.3.

CONCLUSIONS

AAA and MB yield different TCP values as well as NTCP for lung pneumonitis based on the LKB model parameters. Therefore, 2D γ-maps, generated with 2%/2 or 3%/3 mm, could illustrate visual information about the radiobiological changes. The information is useful to evaluate the clinical outcome of a radiotherapy treatment and to approve the treatment plan of the patient if the dose constraints are respected. On the other hand, the γ-maps tool can be used as quality assurance (QA) process to check the predicted TCP and NTCP from radiobiological models.

Radiobiologically optimized couch shift: A new localization paradigm using cone-beam CT for prostate radiotherapy

PURPOSE

To present a novel positioning strategy which optimizes radiation delivery by utilizing radiobiological response knowledge and evaluate its use during prostate external beam radiotherapy.

METHODS

Five patients with low or intermediate risk prostate cancer were evaluated retrospectively in this IRB-approved study. For each patient, a VMAT plan with one 358° arc was generated on the planning CT (PCT) to deliver 78 Gy in 39 fractions. Five representative pretreatment cone beam CTs (CBCT) were selected for each patient. The CBCT images were registered to PCT by a human observer, which consisted of an initial automated registration with three degrees-of-freedom, followed by manual adjustment for agreement at the prostate/rectal wall interface. To determine the optimal treatment position for each CBCT, a search was performed centering on the observer-matched position (OM-position) utilizing a score function based on radiobiological and dosimetric indices (EUDprostate, D99prostate, NTCPrectum, and NTCPbladder) for the prostate, rectum, and bladder. We termed the optimal treatment position the radiobiologically optimized couch shift position (ROCS-position).

RESULTS

The dosimetric indices, averaged over the five patients' treatment plans, were (mean ± SD) 79.5 ± 0.3 Gy (EUDprostate), 78.2 ± 0.4 Gy (D99prostate), 11.1% ± 2.7% (NTCPrectum), and 46.9% ± 7.6% (NTCPbladder). The corresponding values from CBCT at the OM-positions were 79.5 ± 0.6 Gy (EUDprostate), 77.8 ± 0.7 Gy (D99prostate), 12.1% ± 5.6% (NTCPrectum), and 51.6% ± 15.2% (NTCPbladder), respectively. In comparison, from CBCT at the ROCS-positions, the dosimetric indices were 79.5 ± 0.6 Gy (EUDprostate), 77.3 ± 0.6 Gy (D99prostate), 8.0% ± 3.3% (NTCPrectum), and 46.9% ± 15.7% (NTCPbladder). Excessive NTCPrectum was observed on Patient 5 (19.5% ± 6.6%) corresponding to localization at OM-position, compared to the planned value of 11.7%. This was mitigated with radiobiologically optimized localization, resulting in a reduced NTCPrectum value of 11.3% ± 3.5%. Overall, the treatment position optimization resulted in similar target dose coverage with reduced risk to rectum.

CONCLUSIONS

These encouraging results illustrate the potential advantage of applying radiobiologically optimized correction for online image-guided radiotherapy of prostate patients.

Projections onto the Pareto surface in multicriteria radiation therapy optimization

PURPOSE

To eliminate or reduce the error to Pareto optimality that arises in Pareto surface navigation when the Pareto surface is approximated by a small number of plans.

METHODS

The authors propose to project the navigated plan onto the Pareto surface as a postprocessing step to the navigation. The projection attempts to find a Pareto optimal plan that is at least as good as or better than the initial navigated plan with respect to all objective functions. An augmented form of projection is also suggested where dose-volume histogram constraints are used to prevent that the projection causes a violation of some clinical goal. The projections were evaluated with respect to planning for intensity modulated radiation therapy delivered by step-and-shoot and sliding window and spot-scanned intensity modulated proton therapy. Retrospective plans were generated for a prostate and a head and neck case.

RESULTS

The projections led to improved dose conformity and better sparing of organs at risk (OARs) for all three delivery techniques and both patient cases. The mean dose to OARs decreased by 3.1 Gy on average for the unconstrained form of the projection and by 2.0 Gy on average when dose-volume histogram constraints were used. No consistent improvements in target homogeneity were observed.

CONCLUSIONS

There are situations when Pareto navigation leaves room for improvement in OAR sparing and dose conformity, for example, if the approximation of the Pareto surface is coarse or the problem formulation has too permissive constraints. A projection onto the Pareto surface can identify an inaccurate Pareto surface representation and, if necessary, improve the quality of the navigated plan.

Evaluating inter-campus plan consistency using a knowledge based planning model

BACKGROUND AND PURPOSE

We investigate whether knowledge based planning (KBP) can identify systematic variations in intensity modulated radiotherapy (IMRT) plans between multiple campuses of a single institution.

MATERIAL AND METHODS

A KBP model was constructed from 58 prior main campus (MC) esophagus IMRT radiotherapy plans and then applied to 172 previous patient plans across MC and 4 regional sites (RS). The KBP model predicts DVH bands for each organ at risk which were compared to the previously planned DVHs for that patient.

RESULTS

RS1's plans were the least similar to the model with less heart and stomach sparing, and more variation in liver dose, compared to MC. RS2 produced plans most similar to those expected from the model. RS3 plans displayed more variability from the model prediction but overall, the DVHs were no worse than those of MC. RS4 did not present any statistically significant results due to the small sample size (n=11).

CONCLUSIONS

KBP can retrospectively highlight subtle differences in planning practices, even between campuses of the same institution. This information can be used to identify areas needing increased consistency in planning output and subsequently improve consistency and quality of care.

Bulk evaluation and comparison of radiotherapy treatment plans for breast cancer

This study provides a bulk, retrospective analysis of 151 breast and chest wall radiotherapy treatment plans, as a small-scale demonstration of the potential breadth and value of the information that may be obtained from clinical data mining. The treatments were planned at three centres belonging to one organisation over a period of 3 months. All 151 plans were used to evaluate inter-centre consistency and compliance with a local planning protocol. A subset of 79 plans, from one centre, were used in a more detailed evaluation of the effects of anatomical asymmetry on heart and lung dose, the effects of a metallic temporary tissue expander port on dose homogeneity and the overall conformity and homogeneity achieved in routine breast treatment planning. Differences in anatomical structure contouring and nomenclature were identified between the three centres, with all centres showing some non-compliance with the local planning protocol. When evaluated against standard conformity indices, these breast plans performed relatively poorly. However, when evaluated against recommended organ-at-risk tolerances, all evaluated plans performed sufficiently well that tighter planning tolerances could be recommended for future planning. Heart doses calculated in left breast and chest wall treatments were significantly higher than heart doses calculated in right sided breast and chest wall treatments (p < 0.001). In the treatment involving a temporary tissue expander, the inflated implant effectively pushed the targeted breast tissue away from the healthy tissues, leading to a dose distribution that was relatively conformal, although attenuation through the tissue expander's metallic port may have been underestimated by the treatment planning system. The results of this study exemplify the use of bulk treatment planning data to evaluate clinical workloads and inform ongoing treatment planning.

Big Data Analytics for Prostate Radiotherapy

Radiation therapy is a first-line treatment option for localized prostate cancer and radiation-induced normal tissue damage are often the main limiting factor for modern radiotherapy regimens. Conversely, under-dosing of target volumes in an attempt to spare adjacent healthy tissues limits the likelihood of achieving local, long-term control. Thus, the ability to generate personalized data-driven risk profiles for radiotherapy outcomes would provide valuable prognostic information to help guide both clinicians and patients alike. Big data applied to radiation oncology promises to deliver better understanding of outcomes by harvesting and integrating heterogeneous data types, including patient-specific clinical parameters, treatment-related dose-volume metrics, and biological risk factors. When taken together, such variables make up the basis for a multi-dimensional space (the "RadoncSpace") in which the presented modeling techniques search in order to identify significant predictors. Herein, we review outcome modeling and big data-mining techniques for both tumor control and radiotherapy-induced normal tissue effects. We apply many of the presented modeling approaches onto a cohort of hypofractionated prostate cancer patients taking into account different data types and a large heterogeneous mix of physical and biological parameters. Cross-validation techniques are also reviewed for the refinement of the proposed framework architecture and checking individual model performance. We conclude by considering advanced modeling techniques that borrow concepts from big data analytics, such as machine learning and artificial intelligence, before discussing the potential future impact of systems radiobiology approaches.

Estimation of the effects of normal tissue sparing using equivalent uniform dose-based optimization

In this study, we intend to estimate the effects of normal tissue sparing between intensity modulated radiotherapy (IMRT) treatment plans generated with and without a dose volume (DV)-based physical cost function using equivalent uniform dose (EUD). Twenty prostate cancer patients were retrospectively selected for this study. For each patient, two IMRT plans were generated (i) EUD-based optimization with a DV-based physical cost function to control inhomogeneity (EUD_{With DV}) and (ii) EUD-based optimization without a DV-based physical cost function to allow inhomogeneity (EUD_{Without DV}). The generated plans were prescribed a dose of 72 Gy in 36 fractions to planning target volume (PTV). Mean dose, D_30%, and D_5% were evaluated for all organ at risk (OAR). Normal tissue complication probability was also calculated for all OARs using BioSuite software. The average volume of PTV for all patients was 103.02 ± 27 cm³. The PTV mean dose for EUD_{With DV} plans was 73.67 ± 1.7 Gy, whereas for EUD_{Without DV} plans was 80.42 ± 2.7 Gy. It was found that PTV volume receiving dose more than 115% of prescription dose was negligible in EUD_{With DV} plans, whereas it was 28% in EUD_{Without DV} plans. In almost all dosimetric parameters evaluated, dose to OARs in EUD_{With DV} plans was higher than in EUD_{Without DV} plans. Allowing inhomogeneous dose (EUD_{Without DV}) inside the target would achieve better normal tissue sparing compared to homogenous dose distribution (EUD_{With DV}). Hence, this inhomogeneous dose could be intentionally dumped on the high-risk volume to achieve high local control. Therefore, it was concluded that EUD optimized plans offer added advantage of less OAR dose as well as selectively boosting dose to gross tumor volume.

Dosimetric and radiobiological comparison of CyberKnife M6™ InCise multileaf collimator over IRIS™ variable collimator in prostate stereotactic body radiation therapy

The impetus behind our study was to establish a quantitative comparison between the IRIS collimator and the InCise multileaf collimator (MLC) (Accuray Inc. Synnyvale, CA) for prostate stereotactic body radiation therapy (SBRT). Treatment plans for ten prostate cancer patients were performed on MultiPlan™ 5.1.2 treatment planning system utilizing MLC and IRIS for 36.25 Gy in five fractions. To reduce the magnitude of variations between cases, the planning tumor volume (PTV) was defined and outlined for treating prostate gland only, assuming no seminal vesicle or ex-capsule involvement. Evaluation indices of each plan include PTV coverage, conformity index (CI), Paddick's new CI, homogeneity index, and gradient index. Organ at risk (OAR) dose sparing was analyzed by the bladder wall D_max and V_37Gy, rectum D_max and V_36Gy. The radiobiological response was evaluated by tumor control probability and normal tissue complication probability based on equivalent uniform dose. The dose delivery efficiency was evaluated on the basis of planned monitor units (MUs) and the reported treatment time per fraction. Statistical significance was tested using the Wilcoxon signed rank test. The studies indicated that CyberKnife M6™ IRIS and InCise™ MLC produce equivalent SBRT prostate treatment plans in terms of dosimetry, radiobiology, and OAR sparing, except that the MLC plans offer improvement of the dose fall-off gradient by 29% over IRIS. The main advantage of replacing the IRIS collimator with MLC is the improved efficiency, determined from the reduction of MUs by 42%, and a 36% faster delivery time.

Dosimetric and radiobiological impact of intensity modulated proton therapy and RapidArc planning for high-risk prostate cancer with seminal vesicles

INTRODUCTION

The purpose of this study was to evaluate the dosimetric and radiobiological impact of intensity modulated proton therapy (IMPT) and RapidArc planning for high-risk prostate cancer with seminal vesicles.

METHODS

Ten high-risk prostate cancer cases were included in this retrospective study. For each case, IMPT plans were generated using multiple field optimisation (MFO) technique (two fields) with XiO treatment planning system (TPS), whereas RapidArc plans were generated using double-arc technique (two full arcs) with Eclipse TPS. IMPT and RapidArc plans were optimised for a total prescription dose of 79.2 Gy (relative biological effectiveness (RBE)) and 79.2 Gy, respectively, using identical dose-volume constraints. IMPT and RapidArc plans were then normalised such that at least 95% of the planning target volume (PTV) received the prescription dose.

RESULTS

The mean and maximum PTV doses were comparable in IMPT plans (80.1 ± 0.3 Gy (RBE) and 82.6 ± 1.0 Gy (RBE) respectively) and RapidArc plans (80.3 ± 0.3 Gy and 82.8 ± 0.6 Gy respectively) with P = 0.088 and P = 0.499 respectively. The mean doses of the rectum and bladder were found to be significantly lower in IMPT plans (16.9 ± 5.8 Gy (RBE) and 17.5 ± 5.4 Gy (RBE) respectively) when compared to RapidArc plans (41.9 ± 5.7 Gy and 32.5 ± 7.8 Gy respectively) with P < 0.000 and P < 0.000 respectively. For the rectum, IMPT produced lower V₃₀ (21.0 ± 9.6% vs. 68.5 ± 10.0%; P < 0.000), V₅₀ (14.3 ± 5.8% vs. 45.0 ± 10.0%; P < 0.000) and V₇₀ (6.9 ± 3.4% vs. 12.8 ± 3.6%; P < 0.000) compared to RapidArc. For the bladder, IMPT produced lower V₃₀ (23.2 ± 7.0% vs. 50.9 ± 15.6%; P < 0.000) and V₅₀ (16.6 ± 5.4% vs. 25.1 ± 9.6%; P = 0.001), but similar V₇₀ (9.7 ± 3.5% vs. 10.5 ± 4.2%; P = 0.111) compared to RapidArc. RapidArc produced lower mean dose for both the right femoral head (19.5 ± 4.2 Gy vs. 27.4 ± 4.5 Gy (RBE); P < 0.000) and left femoral head (18.0 ± 4.3 Gy vs. 28.0 ± 5.6 Gy (RBE); P < 0.000). Both IMPT and RapidArc produced comparable bladder normal tissue complication probability (NTCP) (0.6 ± 0.2% vs. 0.5 ± 0.2%; P = 0.152). The rectal NTCP was found to be lower using IMPT (0.8 ± 0.7%) than using RapidArc (1.7 ± 0.7%) with P < 0.000.

CONCLUSION

Both IMPT and RapidArc techniques provided comparable mean and maximum PTV doses. For the rectum, IMPT produced better dosimetric results in the low-, medium- and high-dose regions and lower NTCP compared to RapidArc. For the bladder, the NTCP and dosimetric results in the high-dose region were comparable in both sets of plans, whereas IMPT produced better dosimetric results in the low- and medium-dose regions.

Quantification of residual dose estimation error on log file-based patient dose calculation

PURPOSE

The log file-based patient dose estimation includes a residual dose estimation error caused by leaf miscalibration, which cannot be reflected on the estimated dose. The purpose of this study is to determine this residual dose estimation error.

METHODS AND MATERIALS

Modified log files for seven head-and-neck and prostate volumetric modulated arc therapy (VMAT) plans simulating leaf miscalibration were generated by shifting both leaf banks (systematic leaf gap errors: ±2.0, ±1.0, and ±0.5mm in opposite directions and systematic leaf shifts: ±1.0mm in the same direction) using MATLAB-based (MathWorks, Natick, MA) in-house software. The generated modified and non-modified log files were imported back into the treatment planning system and recalculated. Subsequently, the generalized equivalent uniform dose (gEUD) was quantified for the definition of the planning target volume (PTV) and organs at risks.

RESULTS

For MLC leaves calibrated within ±0.5mm, the quantified residual dose estimation errors that obtained from the slope of the linear regression of gEUD changes between non- and modified log file doses per leaf gap are in head-and-neck plans 1.32±0.27% and 0.82±0.17Gy for PTV and spinal cord, respectively, and in prostate plans 1.22±0.36%, 0.95±0.14Gy, and 0.45±0.08Gy for PTV, rectum, and bladder, respectively.

CONCLUSIONS

In this work, we determine the residual dose estimation errors for VMAT delivery using the log file-based patient dose calculation according to the MLC calibration accuracy.

Normal tissue complication probability (NTCP) models for late rectal bleeding, stool frequency and fecal incontinence after radiotherapy in prostate cancer patients

BACKGROUND AND PURPOSE

Curative radiotherapy for prostate cancer may lead to anorectal side effects, including rectal bleeding, fecal incontinence, increased stool frequency and rectal pain. The main objective of this study was to develop multivariable NTCP models for these side effects.

MATERIAL AND METHODS

The study sample was composed of 262 patients with localized or locally advanced prostate cancer (stage T1-3). Anorectal toxicity was prospectively assessed using a standardized follow-up program. Different anatomical subregions within and around the anorectum were delineated. A LASSO logistic regression analysis was used to analyze dose volume effects on toxicity.

RESULTS

In the univariable analysis, rectal bleeding, increase in stool frequency and fecal incontinence were significantly associated with a large number of dosimetric parameters. The collinearity between these predictors was high (VIF>5). In the multivariable model, rectal bleeding was associated with the anorectum (V70) and anticoagulant use, fecal incontinence was associated with the external sphincter (V15) and the iliococcygeal muscle (V55). Finally, increase in stool frequency was associated with the iliococcygeal muscle (V45) and the levator ani (V40). No significant associations were found for rectal pain.

CONCLUSIONS

Different anorectal side effects are associated with different anatomical substructures within and around the anorectum. The dosimetric variables associated with these side effects can be used to optimize radiotherapy treatment planning aiming at prevention of specific side effects and to estimate the benefit of new radiation technologies.

Outcome modeling techniques for prostate cancer radiotherapy: Data, models, and validation

Prostate cancer is a frequently diagnosed malignancy worldwide and radiation therapy is a first-line approach in treating localized as well as locally advanced cases. The limiting factor in modern radiotherapy regimens is dose to normal structures, an excess of which can lead to aberrant radiation-induced toxicities. Conversely, dose reduction to spare adjacent normal structures risks underdosing target volumes and compromising local control. As a result, efforts aimed at predicting the effects of radiotherapy could invaluably optimize patient treatments by mitigating such toxicities and simultaneously maximizing biochemical control. In this work, we review the types of data, frameworks and techniques used for prostate radiotherapy outcome modeling. Consideration is given to clinical and dose-volume metrics, such as those amassed by the QUANTEC initiative, and also to newer methods for the integration of biological and genetic factors to improve prediction performance. We furthermore highlight trends in machine learning that may help to elucidate the complex pathophysiological mechanisms of tumor control and radiation-induced normal tissue side effects.

Bayesian network ensemble as a multivariate strategy to predict radiation pneumonitis risk

PURPOSE

Prediction of radiation pneumonitis (RP) has been shown to be challenging due to the involvement of a variety of factors including dose-volume metrics and radiosensitivity biomarkers. Some of these factors are highly correlated and might affect prediction results when combined. Bayesian network (BN) provides a probabilistic framework to represent variable dependencies in a directed acyclic graph. The aim of this study is to integrate the BN framework and a systems' biology approach to detect possible interactions among RP risk factors and exploit these relationships to enhance both the understanding and prediction of RP.

METHODS

The authors studied 54 nonsmall-cell lung cancer patients who received curative 3D-conformal radiotherapy. Nineteen RP events were observed (common toxicity criteria for adverse events grade 2 or higher). Serum concentration of the following four candidate biomarkers were measured at baseline and midtreatment: alpha-2-macroglobulin, angiotensin converting enzyme (ACE), transforming growth factor, interleukin-6. Dose-volumetric and clinical parameters were also included as covariates. Feature selection was performed using a Markov blanket approach based on the Koller-Sahami filter. The Markov chain Monte Carlo technique estimated the posterior distribution of BN graphs built from the observed data of the selected variables and causality constraints. RP probability was estimated using a limited number of high posterior graphs (ensemble) and was averaged for the final RP estimate using Bayes' rule. A resampling method based on bootstrapping was applied to model training and validation in order to control under- and overfit pitfalls.

RESULTS

RP prediction power of the BN ensemble approach reached its optimum at a size of 200. The optimized performance of the BN model recorded an area under the receiver operating characteristic curve (AUC) of 0.83, which was significantly higher than multivariate logistic regression (0.77), mean heart dose (0.69), and a pre-to-midtreatment change in ACE (0.66). When RP prediction was made only with pretreatment information, the AUC ranged from 0.76 to 0.81 depending on the ensemble size. Bootstrap validation of graph features in the ensemble quantified confidence of association between variables in the graphs where ten interactions were statistically significant.

CONCLUSIONS

The presented BN methodology provides the flexibility to model hierarchical interactions between RP covariates, which is applied to probabilistic inference on RP. The authors' preliminary results demonstrate that such framework combined with an ensemble method can possibly improve prediction of RP under real-life clinical circumstances such as missing data or treatment plan adaptation.