Dosimetric Evaluation of 3-Dimensional Conformal Radiotherapy Technique in Postoperative Patients with Gastric Carcinoma: When Is IMRT Really Needed?

Stomach cancer is the fifth most common cancer in the world (5.7% of all new cases) and the third most common cause of cancer related death (8.2% of all cancer deaths) [1]. India has a low incidence of stomach cancer compared to world statistics and is the fifth most common cancer in India. The estimated incidence of gastric cancer in India for the year 2018 was 5% (57,394 cases) and about 38,818 new gastric cancer cases were estimated to have occurred in males [2]. Partial or total gastrectomy is the mainstay of treatment in the management of stomach cancer [3]. In the majority of cases, treatment with surgery alone is not satisfactory and the 5-years survival depends Abstract

Introduction the present study was to do a dosimetric analysis of the doses received in post-operative setting by using 3D-CRT to a total dose of 45 Gy in 25 fractions over five weeks with concurrent chemotherapy. A subsequent comparison with the existing 3D-CRT and IMRT literature was done.

Materials and Methods
This retrospective analysis on dosimetric data in post-operative carcinoma stomach was carried out at the Department of Radiotherapy and Oncology in a multispecialty tertiary care centre in India after informed consent. Being a retrospective dosimetric analysis of patients treated as per institutional protocol, additional ethical clearance was not sought. Case records of all patients with carcinoma stomach (adenocarcinoma and signet ring cell carcinoma) registered in the department between January 2012 and December 2015 were analysed. Out of 156 cases of primary gastric cancer, only 93 (60%) of the cases were histopathologically confirmed to have adenocarcinoma and signet ring cell carcinoma. Out of these, the 60 patients who underwent radical surgery followed by post -operative radiotherapy have been included in this analysis and the rest who received palliative treatment were omitted. All patients received concurrent chemotherapy using 5-FU or capecitabine based regimen. Staging was done according to AJCC 7 th edition staging system [10].

Radiotherapy Planning
The post-operative patients of carcinoma of stomach were selected for this dosimetric analysis of the doses received by PTV and OARs using 3D-CRT. All patients underwent total or subtotal gastrectomy with D1 or D2 lymph-nodal dissection. Planning CT scans were taken with patients in supine position using multislice CT scanner with a slice thickness of 2.5mm (GE healthcare technologies, Wankesha, WI, USA). The images were then transferred to EclipseTM treatment planning system (v.8.6, Varian associates, Palo, Alto, CA, USA). Contouring was done as per Tepper and Gunderson guidelines [10][11]. The clinical target volume (CTV) included both post-operative tumour bed and regional lymphatic drainage. The nodal CTV included nodes in the para-aortic (PA), celiac axis, porta hepatis and superior mesenteric artery (SMA) group. All lymph nodal groups were contoured according to the guidelines be Tepper et al [10]. The PA lymph-nodal volume was delineated by giving differential margin to the aorta from thoracic (T11) to lumbar (L2) vertebrae, (i.e. 2.5 cm right side, 2 cm anterior, 1.2 cm left side, and 0.2 cm posterior), 1cm symmetric margin was given to portal vein (from origin of portal vein to porta-hepatis), celiac artery (from origin of celiac axis to its first bifurcation) and superior mesenteric artery (from its origin to 1 cm) ( Figure 1). The CTV tumour included the tumour bed, remaining stomach and anastomotic site. Both CTV nodal and CTV tumour were dependent on the primary location of the tumour, pathological T-stage and nodal stage. The CTV total was made by joining both CTV nodal and tumour with the help of Boolean apparatus. PTV was contoured by giving 1 cm symmetric margin around CTV total. OARs including liver, both kidneys, spinal cord, heart, and small bowel were contoured. 3D-CRT plans were generated by using one anterior and two lateral fields with appropriate wedges (Figure 2) with 6 MV photons and treatment was executed by a linear accelerator. A total dose of 45 Gy in 25 fractions were prescribed to the PTV. Plans were optimised whenever required to deliver the prescribed dose to more than 95% of PTV. Dose volume histograms (DVH) were generated and evaluated for all the OARs before delivering the treatment. Evaluation of dosimetric data of the target volumes and OARs was done using Quantitative Analysis of Normal Tissue Effects in Clinics (QUANTEC) parameters [12].
All 3D-CRT plans were analysed in terms of PTV coverage, conformity index (CI), homogeneity index (HI), and OAR dose volume parameters. Several definitions of CI and HI are available but we used the RTOG definitions to calculate the CI and HI.
Conformity index (RTOG definition) = Reference isodose volume / target volume Homogeneity index = maximum isodose in the target / reference isodose In an ideal scenario the CI should be equal to 1 and HI should be < 2. CI less than 1 indicates that the target volume is not adequately irradiated and a value greater than 1 signifies that the irradiated volume is greater than the target volume.

Statistics
Statistical analysis was done with the Statistical Program for Social Sciences (SPSS v23, IBM Corp, USA). Descriptive analysis was done for the dosimetric and demographic data. Summary of statistics including mean, range, and standard deviation were obtained.

Patient and treatment characteristics
The median age of patients in our study was 52 years (22-75 years). Most of them were males (33 patients, 55%). Most of the patients were smokers (60%) and occasional

Discussion
Partial or total gastrectomy along with the D1 or D2 lymph-node dissection is the mainstay of treatment in the management of stomach cancer. Loco-regional recurrence is the major problem in patients treated with surgery alone necessitating various attempts at improving survival using alcoholics (55%). Five percent of the patients had a family history of malignancy. The common presenting symptoms were loss of appetite, upper abdominal pain and loss of weight. Most of the patients underwent sub-total gastrectomy (90%) with D1 lymph-node dissection (66%) or D2 lymph-node dissection (33%). Adenocarcinoma (intestinal type) was the most common post-operative histopathological variant. All patients underwent post-operative radiotherapy to a total dose of 45 Gy in 25 fractions over 5 weeks using 3D-CRT (Table 1).
With the introduction of 3D-CRT, dose to the OARs could be minimized with adequate delivery of dose to the tumour bed and nodal basin [4][5][6][7][8][9][10][11][12]. In our study, by using the standard RTOG tools, both HI and CI were within the range of reported IMRT data [13]. In a study by Li Z et al, the author compared dosimetric parameters of 5 field intensity-modulated radiotherapy (5F-IMRT), 7F-IMRT, single-arc volumetric modulated arc therapy (SA-VMAT) and double arc (DA)-VMAT [12]. All plans were made with the prescription dose of 50.4 Gy in 28 fractions. CI was 0.86±0.02 in both 5F-IMRT and 7F-IMRT planning's, 0.83±0.03 for SA-VMAT, and 0.87±0.03 for DA-VMAT planning. CI was better with the IMRT and ARC therapy compared to 3D-CRT planning in the current study due to the forward planning of 3D-CRT.
Kidney is a radiosensitive organ and abdominal or pelvic radiotherapy leads to damage to one or both the kidneys [14]. For bilateral kidneys, V20 <32%, mean dose < 15-18Gy, V12 <55% is recommended to get <5% of radiation induced kidney damage [15][16]. In the present study, mean dose to the left kidney was higher than the right kidney, the primary reason being that the left kidney in most of the patients was located within the target volume and tumor bed was anterior to the left kidney. However, the doses were comparable to that reported in various studies as outlined in Table 4 by both IMRT and 3D-CRT. Therefore, location of the post-operative bed rather than the technique appeared to be the primary limiting factor. Alani et al. defined separate high and low dose kidney instead of laterality to incorporate this concept.
Liver is a parallel organ, and the radiation induced liver damage (RILD) is found to correlate with the volume and dose of radiation received by the normal liver tissue. The incidence of RILD is dependent on the mean dose to the normal liver tissue with models predicting no cases of RILD with a mean liver dose < 31 Gy [20]. In the present study, mean dose to the liver was 27.7±6.4 Gy and V20 was 69.2±15.8 which were within the tolerable limit for the normal liver tissue. Though mean doses to the liver were lower with IMRT as reported by Li Z et al [12] and Minn et al. [8], 3D-CRT planning yielded values well within tolerance limits.
Severe gastro-intestinal (GI) toxicities are the main dose limiting factor in case of abdominal irradiation. In a comparison study of IMRT and 3D-CRT for adjuvant treatment of gastric cancer by Minn AY et al, grade 2 or higher GI toxicity was found to be similar between IMRT and 3D-CRT patients (61.2% and 61.5% respectively) and V45 for the bowel space is 106.2 cc, Dmax is 50.5 Gy for IMRT planning [18]. In another study by Liu GF et al, acute GI toxicities were found to be 56% and 54% for IMRT and 3D-CRT respectively [21]. Li Z et al showed that the Dmean for small intestine was 17.19±6.26 Gy for 5F-IMRT, 16   In the present study, Dmax for the bowel bag is 48.7±1.4 Gy, V45 is 29.2±9.8%, and D195cc is 36.3±6.7 Gy which showed that the doses were comparable to the above mentioned studies. For individual patients more than the radiotherapy technique it was the position of the anastomotic site which influenced high OAR doses (kidney and liver). The anastomotic sites and DVHs comparing kidney doses of an individual patient are shown in Figure 4.
Limitation of the current study are its retrospective nature and there is no direct comparison of 3D-CRT planning with the IMRT planning for the same set of patients.
In conclusion, results of the study suggest that there is a challenge in achieving the Organ at Risk (OAR) constraints in respect to bowel bag and mean doses achieved for left kidney in patients who had undergone D1 lymphadenectomy. It was also observed that inclusion of postoperative site was having changing position depending upon patient to patient which further led to difficulty in achieving constraints. Some of these patients may favour IMRT based treatment to have better dose delivery.
In the current study we found that post-operative 3D-CRT in cases of carcinoma stomach is acceptable in terms of target volume coverage and homogeneity.
The strength of our current study is that it contains detailed dosimetric analysis in terms of physical parameters of target volume, HI, CI and OAR doses in a homogenous group of post-operative cases of carcinoma of stomach. In a country like India where we have increased patient load, we can spend less time on machine using 3D-CRT. Based on the result we could conclude that 3D-CRT offers good coverage of target volumes at the same time sparing the critical surrounding organs.