Comparative study of automatic and manual planning methods for volumetric modulated arc therapy in patients with intraocular cancer

To research and assess automatic volumetric modulated arc therapy (VMAT) planning methods for patients with intraocular cancer.

because of its lower rates of late recurrence compared with radioactive plaque brachytherapy. 4 EBRT also has the advantage of preserving the eye structure, which results in a better appearance after treatment compared with surgery. 5 One of the major concerns of EBRT is the unnecessary radiation dose delivered to the surrounding normal tissues, which might result in cataracts, retinal detachment, glaucoma, or bleeding into the eye, among other complications. 6,7 Breakthroughs in EBRT technologies, such as proton therapy, intensity-modulated radiation therapy (IMRT), and volumetric modulated arc therapy (VMAT), have been widely investigated to reduce dose delivery to normal tissues while ensuring tumor coverage. 8,9 The possibility and advantages of VMAT for the treatment of intraocular cancer were assessed in one of our previous investigations, and it yielded promising results. 10 However, one of the disadvantages of VMAT is the long optimization time required, which is usually four-to fivefold that of IMRT. 11 Furthermore, a significant interand intra-institutional variation in planning practice and quality were observed for IMRT and VMAT. 12 The present study aimed to investigate the feasibility and advantages of knowledge-based automatic VMAT planning for the treatment of patients with intraocular cancer.
Furthermore, this investigation responds to the increasing demand for automatic VMAT planning for intraocular cancer.

Plan evaluation
The automatic VMAT plans were evaluated according to standard dose-volume histogram parameters. D99% and D1% (dose irradiated to 99% and 1% volume) were reported as a measure of the minimum and maximum doses for PTV, and V95% (volume receiving at least 95% of the prescribed dose) defined as the target coverage was reported. The Homogeneity Index (HI), which was defined as the difference between D1 and D99 divided by prescription dose (Dp), was calculated according to the formula shown below, and reported 15 : An equivalent uniform dose (EUD) was also calculated and reported for the analysis of radiobiological effects of physics dose resulting from Where N is the number of voxels in the structure of interest, Di is the dose in the i-th voxel, and a is a specific parameter of the normal tumor tissue describing the dose-volume effect. The machine unit of the manual and automatic planning, as well as time of delivery, were also evaluated and compared.

Statistical analysis
The evaluation results were described as the mean ± standard deviation. A comparison of the dosimetric and non-dosimetric indices between the two treatment plans was analyzed using the Wilcoxon signed-rank test. All statistical analyses were carried out using SPSS software (version 20.0.0; IBM Corporation, Armonk, NY, USA). Differences were considered statistically significant when P < 0.05.

RESULTS
A total of 10 patients (5 men and 5 women) with primary and secondary intraocular cancer were enrolled in this study, with a median age of 52 years (range 33-80 years). Most patients were diagnosed as having mucosa-associated lymphoid tissue lymphoma. The PTVs of these ated to the ipsilateral lens and optical nerves, but it also increased the dose delivered to the brainstem compared with manual planning, as shown in Table 3. No significant differences in contralateral lens, eyeballs, or optic chiasm were seen as a consequence of either plan-

DISCUSSION
In the present study, we investigated automatic planning for VMAT in the treatment of intraocular cancer. Automatic planning achieved acceptable target coverage and sparing of OARs compared with manual planning. Nevertheless, the add-in algorithm greatly increased optimization time.
Enucleation, RT, cryotherapy, and laser therapy are common local therapy options for the management of intraocular cancer. 1 Lymphoma F I G U R E 1 Typical dosimetric distribution and comparison between automatic and manual volumetric modulated arc therapy plans F I G U R E 2 Typical dose-volume histogram comparison between automatic and manual volumetric modulated arc therapy plans accounts for a major portion of patients with intraocular cancer, and One of the limitations of knowledge-based automatic planning is the limitation in selecting the training dataset, which is randomly selected from the clinical plans used for previous patient treatments. 22,23 The quality of the automatic plans depends on the quality of the plans in the training dataset. Relatively few plans were enrolled in the present study's training set due to the limited cases of intraocular cancer.
Another limitation of this study was the relatively longer optimization time required for automatic planning as an added-in algorithm. One objective of automatic planning is to decrease planning time and solve the shortage of sufficiently trained personnel. 24 However, automatic planning time was much longer than manual planning in the present study. Further investigation of increased treatment planning system capacity and performance is required to explore the benefits of automatic VMAT plans for intraocular cancer.
Automatic VMAT planning achieved acceptable target coverage and lowered the dose delivered to the ipsilateral lens and brainstem for patients with intraocular cancer in comparison with manual planning. Automatic planning is a promising method for the management of intraocular cancer; however, the full benefits of automatic VMAT planning require further investigation for improved performance and capacity of the treatment planning system.