Patient selection and planning
We randomly selected 13 local patients with rNPC who underwent CIRT at our hospital from June 2016 to December 2017.
Target definition (4): The gross tumor volume (GTV) includes visible tumor lesions on CT, PET-CT, and MRI. The clinical target volumes (CTV1) of both the GTV of the primary site and neck were designed to include 5 mm beyond the GTV for microscopic extension (limited to as little as 1 mm near OAR), and a variable margin for occult tumor spread. CTV2 includes CTV1 and subclinical lesions that may be invaded by the tumor. The planning target volume (PTV) is based on a CTV expansion of 6 mm in the direction lateral to the beams and 3 mm in other directions, which is calculated based on the range uncertainty (19) and allows for setup variability and uncertainty about dose distribution. Optimizing PTV can help meet target dose requirements.
Prescription: Nine of the patients with recurrent disease had a CTV1 prescription of 63.00 Gy (RBE) / 21 fractions, while the remaining four patients had a CTV1 prescription of 60.00 Gy (RBE) / 20 fractions. Their dose per fraction was the same 3.00 Gy (RBE). Since the dose conversion is only related to the dose per fractions (13), the total fractionations of all patients were rescaled to 21 fractions in the new treatment plans.
This study was performed using the Raystation (V8A, Raysearch, Sweden) treatment planning system, which incorporates both MKM and LEM. The RBE-weighted doses calculated by LEM and MKM are hereafter referred to as the LEM dose and MKM dose. Selected patients were used to first generate LEM plans based on LEM. The plan pass criteria (4) of the LEM plan were: CTV coverage requires 99% CTV to cover at least 95% of the prescribed dose (V95 > 99%); maximum brain stem dose (Dmax) < 45.00 Gy (RBE); maximum spinal cord dose < 30.00 Gy (RBE); 20% optic nerve / chiasma received dose D20 < 30.00 Gy (RBE). Next, MKM (10) was used to recalculate the physical doses obtained from the optimization of the LEM plan and created corresponding MKM plans to obtain the MKM dose distributions.
Isovolumetric Dose Method
Wang (18) has analyzed the feasibility of RBE-weighted dose conversion from MKM to LEM. Hence, if the physical dose and fragment spectrum are exactly the same, the LEM isodose can be transformed into a defined MKM isodose. Also, under the clinical treatment plan, they have the same biological equivalent dose-volume, which is a more efficient tool to establish the conversion relationship.
Conversion Curve
Previous scholars (15) defined the conversion factor as the ratio of the LEM dose to the MKM dose. For the dose conversion inside CTV, we directly focused on the dose in the target volume of CTV1 and CTV2 in the LEM and MKM plan of each patient.
For the dose conversion outside CTV, we first defined the dose area of interest outside the CTV as the CTV 20 mm extension (exclude CTV), which includes all the OAR adjacent to the CTV. Then, 56 isodose curves of 60.00 Gy (RBE) to 5.00 Gy (RBE) were selected in the LEM plans of 13 patients, whose fractional doses ranged from 2.86 Gy (RBE) to 0.24 Gy (RBE). The volume of each isodose line was obtained, and then the corresponding MKM dose of the same volume was found in the MKM plan. A series of conversion factors were obtained according to the definition.
The conversion of the OAR constraints
Most patients with head and neck tumors in NIRS received 16 fractions of radiation, while the methods of this study used 21 fractions, as described above. Since the total dose in multi-fraction irradiations depends more on the size of dose-per-fraction for late, rather than for early, damage to normal tissues (20), the Linear-quadratic (LQ) model should be first used to convert the dose limits in 21 fractions of the LEM plan to the dose limits in 16 fractions. Under the same fractionations, a single MKM dose can be obtained by using the conversion curve in this study. Finally, we multiplied the single dose of MKM by 16 to obtain the total corresponding MKM dose for the 16 fractions.