Side effects of CT-guided implantation of ¹²⁵I seeds for recurrent malignant tumors of the head and neck assisted by 3D printing non co-planar template

Forty-two patients with recurrent/metastasis of a malignant tumor of the head and neck who received RIS implantation assisted by a 3D–PNCT and CT guidance in our hospital from January 2016 to October 2016 were enrolled. Among them, 26 cases had recurrence of a primary tumor of the head and neck and 16 cases had a metastatic tumor of the neck. Also, 64.3% (27/42) of the patients had undergone head and neck surgery, 85.7% (36/42) of cases had accepted EBRT, and 69% (29/42) of cases had a history of chemotherapy. The Karnovsky Performance Status (KPS) score of all patients was > 70. General information of the patients is shown in Table 1. According to the literature and clinical experiences within our center, 110–160 Gy was the recommended for prescription doses [3, 5, 7, 8, 13‐19].

All patients underwent spiral CT (Brilliance Big Bore; Philips, Amsterdam, the Netherlands) 2 days before surgery. Patients were positioned prone or supine position according to the tumor site. Then, patients were fixed with vacuum pads and marked with a positioning line on the body surface. CT data were transmitted to a Brachytherapy Treatment Planning System (B-TPS). A RIS implantation planning system was designed by the Imaging Center of the Beijing University of Aeronautics and Astronautics (Beijing, China) for preoperative planning. The designed treatment plan on 2D and 3D CT images involved: (i) delineation of the gross tumor volume (GTV) and adjacent organs at risk (OARs); (ii) setting of the prescribed dose and RIS activity; (iii) determination of the needle tract of the implanted RIS (insertion direction, distribution, and depth); (iv) calculation of the RIS number and simulation of the spatial distribution of RIS; (v) calculation of the dose distribution of the target volume and OARs (parotid gland, spinal cord, trachea, mucous membrane, skin, important blood vessels in the neck). We optimized the plan to make the doses of 90% GTV (D90 of GTV) reach the prescribed doses (110–150 Gy) as far as possible while ensuring that the exposure dose to OARs was as low as possible.

Depending on B-TPS data, we used 3D images and reverse-engineering software to establish a digital model for the individual template. Coordinates for the spatial points of RIS and needle-tract orientation were incorporated into the model. Then, we used 3D rapid-prototyping equipment with photo-curable resins to print individualized 3D templates. The template contained the superficial anatomic characteristics of the treatment area, positioning markers and simulation of the needle tract.

Local anesthesia was induced in all patients. The 3D–PNCT was placed on the surface of the treatment area of the patient. The 3D–PNCT was aligned accurately with the outer-contour features, positioning marks, and positioning laser line. Through the guide hole of the 3D–PNCT, we inserted 3 stable needles (Mick Radio Nuclear Instruments, Mount Vernon, NY, USA) percutaneously to pre-planned depths. We carried out CT immediately to validate the locations of stable needles and, if necessary, fine-tuned the intraoperative plan in real time. Finally, depending on preoperative and intraoperative planning, we implanted the RIS in a retrusive manner with a Mick gun (Mick Radio Nuclear Instruments of the USA). Iodine-125 seeds (Tian Jin Sai De Tech Co., Ltd. Atom High Tech Co., Ltd. Chinese Tong Fu Tech Co., Ltd.) with half life of 59.4 days and dose rate constant of 0.965 cGy/(h·U). CT was undertaken immediately after completion of RIS implantation to observe the actual distribution of RIS (Fig. 1).

Dosimetric evaluations of GTV and adjacent OARs were based on postoperative CT images.

Tumor response was evaluated at 4 weeks and every 2–3 months thereafter. Follow-up involved regular visits and telephone conversations.

The side effects of skin puncture were bleeding, pain, infection, non-union of puncture point, and metastasis due to RIS implantation. The side effects of radiation (which were scored using the toxicity criteria of the Radiation Therapy Oncology Group (RTOG) and the European Organization for Research and Treatment of Cancer (EORTC) were skin injury, mucosal response, spinal-cord injury, peripheral-nerve injury, xerostomia and blood toxicity [20]. Nerve injury was scored by the Common Terminology Criteria for Adverse Events v4.0 [21]. Acute toxicities were evaluated in the first 2 months, and late toxicity were avaluated in 2 months later.

The seed migration, one of specific complications of RIS implantation, was observed during the follow-up period.

With regard to side effects, the chi-squared test was used to analyze significant differences among the grades for the same factor. Related factors were the dose absorbed by the OARs, previous accumulative dose, and preoperative physical condition of the OARs.

All forty-two patients completed CT-guided RIS implantation assisted by a 3D–PNCT. The activity of a single RIS was 0.34–0.7 (median, 0.6) mCi. The number of RIS was 10–126 (median, 34). The number of implantation needles was 4–31 (median, 11). All but one episodes of intraoperative pain were tolerable well and disappeared after the surgical procedure. One case had severe intraoperative pain that was barely tolerable, but was relieved postoperatively. Intraoperative bleeding was not obvious. No patients suffered infection at the puncture site after RIS implantation. Skin healing at the puncture site was good except for two cases in which the healing time was delayed for ≈1 week due to exudation. Metastasis due to RIS implantation was not found during follow-up.

All the patients were followed up, and the follow-up rate was 100%. The follow-up time was 4 ~ 14 months (median, 8.5 months).

No case had an acute reaction of grade ≥ 3. Three cases had a grade-1 skin reaction: one case had a grade-1 mucosal reaction and two cases a grade-2 reaction. Blood toxicity was not observed.

With regard to late responses, spinal-cord injury did not occur. Xerostomia was not aggravated. One case had a grade-3 nerve response. Preoperatively, this patient had an abnormal sensation with tumor invasion of the brachial plexus and weak lifting force (strength was level 4). Three months after brachytherapy, the tumor had nearly disappeared, but the neurologic symptoms were aggravated, the abnormal sensation was more obvious than before, and the lifting force was weaker (level 3). The nerve reaction worsened from level 2 (preoperative) to 3 (postoperative) (Common Terminology Criteria Adverse Events v4.0). Thus, we speculated that it was more likely to be a radiation-induced nerve injury. However, no evidence could exclude the possibility of tumor invasion because irreversible nerve injury had occurred preoperatively which worsened postoperatively. The severity of the skin reaction was related to the absorbed dose (D0.1cc) (P = 0.011). There was little evidence that skin edema and fibrosis were related to the accumulative dose of previous EBRT. One case had nerve injury which might have been related to tumor involvement on nerves. RIS translocation did not occur (Tables 2, 3 and 4).