Association between tumor morphology and dosimetric parameters of organs at risk after intensity‐modulated radiotherapy in esophagus cancer

PURPOSE: We explored the effects of geometrical topological properties of tumors such as tumor length and “axial cross‐sectional area (ACSA)” of tumors (planning target volume [PTV] volume /PTV length) on the dosimetric parameters of organs at risk (lung and heart) in patients with esophagus cancer (EPC) treated by way of intensity‐modulated radiation therapy (IMRT), so as to provide a guideline for the dosimetric limitation for organs at risk in IMRT treatment. METHODS: A retrospective analysis was done on 103 cases of patients with EPC who were treated by IMRT from November 2010 to August 2019, in which PTV‐G stood for the externally expanded planning target volume (PTV) of the gross tumor volume (GTV) and PTV‐C for the externally expanded volume of the clinical target volume (CTV). A linear regression model was employed to analyze the several pairs of correlation: the 1st one between the relative length of tumors (PTV length/lung length) and pulmonary dose‐volume parameters, the 2nd one between ACSA of tumors and pulmonary dose‐volume parameters, the 3rd one between PTV length and the dosimetric parameters of the heart, and the last one between ACSA of tumors and the dosimetric parameters of the heart. RESULTS: (i) There was a strong positive correlation between the relative length of tumors (PTV length/lung length) and V (5) (p < 0.001, r = 0.73), and V (10) (p < 0.001, r = 0.66) of the lung. There was a moderate positive correlation between the relative length of tumors and V (30) (p < 0.001, r = 0.44) of the lung, and a weak positive correlation between the relative length of tumors and V (20) (p < 0.001, r = 0.39) of the lung. (ii) There was a strong positive correlation between ACSA of tumors (PTV volume/PTV length) and V (30) (p < 0.001, r = 0.67) of the lung, a moderate positive correlation between ACSA of tumors and V (20) (p <0.001, r = 0.51) of the lung, and a weak positive correlation between ACSA of tumors and V (10) (p = 0.019, r = 0.23) of the lung, yet there was not an obvious correlation between ACSA of tumors and V (5) p > 0.05) of the lung. (iii) There was a moderate positive correlation between PTV length and V (40) (p < 0.001, r = 0.58), and D (mean) (p < 0.001, r = 0.52) of the heart, yet there was no obvious correlation between ACSA of tumors and D (mean) and V (40) of the heart (p > 0.05). CONCLUSIONS: (i) Compared with the high‐dose region of the lung, the relative length of tumors (PTV length/lung length) has a greater impact on the low‐dose region of the lung. The linear regression equation of scatter plot showed that when the relative length of tumors increased by 0.1, the lung dose‐volume parameters of V (5), V (10), V (20), and V (30) increased by approximately 5.37%, 3.59%, 1.05%, and 1.08%, respectively. When PTV length increased by 1 cm, D (mean) and V (40) of the heart increased by approximately 153.6 cGy and 2.03%, respectively. (ii) Compared with the low‐dose region of the lung, the value of ACSA of tumors (PTV volume/PTV length) has a greater impact on the high‐dose region of the lung. However, the value of ACSA of tumors has no significant effect on the dosimetric parameters of the heart (D (mean) and V (40)). The linear regression equation of scatter plot showed that when ACSA of tumors increased by 10 cm(2), the lung dose‐volume parameters of V (10), V (20,) and V (30) increased by approximately 3.11%, 3.37%, and 4.01%, respectively.