Missed radiation therapy sessions in first three weeks predict distant metastasis and less favorable outcomes in surgically treated patients with oral cavity squamous cell carcinoma

We retrospectively reviewed the clinical records of 1058 adult patients with histology-proven OCSCC who received radical surgery and RT – either alone or in combination with chemotherapy – at the Linkou Chang Gung Memorial Hospital (Taoyuan City, Taiwan) between January 2005 and December 2012. All participants were restaged according to the American Joint Committee on Cancer (AJCC) Staging Manual, Eighth Edition by an experienced radiation oncologist after reviewing medical records. The institutional tumor board performed central review of each new case. Before the routine use of positron emission tomography (PET) imaging (2009), staging was performed with computed tomography (CT) or magnetic resonance (MR) imaging along with abdominal ultrasound and bone scan. Currently, both MR and PET can be performed for staging purposes. As of 2009, tumor board routinely suggested a second PET scan (post-operative/pre-RT PET) prior to adjuvant treatment for high-risk patients (e.g., those presenting with extranodal spread) to detect early recurrences. Only patients who underwent surgery with curative intent (including neck dissection) and who received a radiation dose of 2 Gy per fraction were deemed eligible. Exclusion criteria were as follows: 1) history of second primary malignancies; 2) local, regional, or distant failure occurring during the course of RT; 3) total RT dose < 60 Gy or > 72 Gy; and 4) presence of distant recurrences identified on the second PET scan. Owing to the retrospective nature of the study, the need for informed consent was waived.

Primary tumors were excised with ≥1 cm margins (both peripheral and deep margins). Patients with clinically positive nodal disease underwent level I–V neck dissections, whereas level I–III neck dissections were performed in presence of clinically negative nodes. The procedures for collection and classification of pathologic risk factors and the indications for adjuvant treatment have been previously described in detail [16, 17]. Patients received homogeneous treatment according to our institutional guidelines (Supplementary Table 1). All patients underwent postoperative radiotherapy (PORT) consisting of a conventional fractionated daily dose of 2 Gy per fraction, 5 days per week. A 6-MV photon beam was used to achieve a target total dose between 60 and 66 Gy. Suspicious FDG-avid lesions detected on PET imaging before the beginning of RT received a simultaneous integrated boost at a dose of 70–72 Gy [18]. The initial treatment volume comprised the primary tumor bed and the regional cervical nodes. PORT was performed using different techniques – including conventional two-dimensional RT, three-dimensional conformal RT, intensity-modulated radiation therapy (IMRT), and volumetric-modulated arc therapy (VMAT). A conventional field arrangement generally included a bilateral opposing field and a low anterior portal – with the spinal cord being shielded upon delivering of a dose of 46 Gy or more. The dose delivered to the brain stem and spinal cord was limited to 54 Gy. Upon administration of 46–50 Gy, the irradiation area was reduced to include the tumor bed and metastatic nodes only. Concurrent chemotherapy was offered to patients harboring adverse prognostic factors [17]. Cisplatin – generally administered as a single dose of 100 mg/m² every 3 weeks or at a weekly dose of 40 mg/m² – was the most commonly used chemotherapy agent [19, 20].

The performance status was calculated with the Eastern Cooperative Oncology Group scale. Cigarette smoking was dichotomized as yes (subjects who smoked ≥100 cigarettes in their lifetime) versus no (subjects who smoked < 100 cigarettes in their lifetime and who were not currently smoking) [21]. Alcohol consumption (current or former drinkers versus nondrinkers) and betel quid chewing (current or former chewers versus non-chewers) were also considered as dichotomous variables. Chemotherapy was dichotomized as yes (concurrent chemotherapy or chemotherapy administered in the 2 weeks preceding the start of RT) versus no. The Charlson Comorbidity Index [22] was used to categorize the presence of comorbidities as yes (score ≥ 1) versus no. Preoperative PET imaging was dichotomized as yes versus no.

Because a prolonged RT course is known to be associated with a less favorable OS [23], we solely focused on patients with a radiation treatment time (RTT) – defined as the time from the beginning to the end of RT – of less than 8 weeks. The treatment package time (TPT) was calculated from the date of surgery to the last RT session. The interval between surgery and RT was calculated from the date of surgery to the first RT session. For the purpose of analysis, the study participants were divided into three groups based on the characteristics of missed RT, as follows: 1) early missed RT (defined as having received < 14 fractions in the first 3 weeks), 2) late missed RT (defined as having received ≥14 fractions in the first 3 weeks but less than 29 fractions in the first 6 weeks), and 3) RT as scheduled (defined as having received ≥14 fractions in the first 3 weeks and ≥ 29 fractions in the first 6 weeks).

OS – calculated as the time elapsed (in years) from the start of RT to the date of death – was the main outcome measure. Secondary outcomes included LRC and FFDM rates. LRC was defined as the time from the start of RT to the date of local or regional recurrence, whereas FFDM was the time elapsed from the start of RT to the date of diagnosis of distant metastases.

Intergroup differences in terms of continuous variables were assessed with the Student’s t-test, whereas the chi-square test was used for categorical data. Survival curves were plotted with the Kaplan-Meier method (log-rank test). Multivariate Cox proportional hazards regression analyses were used to identify independent predictors of OS, LRC, and FFDM rates. The following covariates were entered into the multivariate model: age, sex, pT, pN, tumor differentiation, TPT, pattern of missed RT sessions (early missed RT, late missed RT, RT as scheduled), cigarette smoking, betel quid chewing, alcohol consumption, presence of comorbidities, concurrent chemotherapy, and PET imaging. The results are expressed as hazard ratios (HRs) with 95% confidence intervals (CIs). In all analyses, two-tailed p values < 0.05 were considered statistically significant.

The median age of the 905 study participants was 50.8 years (range, 25.1–89.4 years) (Table 1). There were 37 (3.1%) patients who had pathological stage I − II disease, whereas 868 (95.9%) had stage III − IV disease. Pathological nodal metastases were identified in 409 patients (63.3%). PET/CT imaging was performed during the preoperative staging work-up in 442 (48.8%) study participants. Concurrent chemoradiation was administered to 505 (55.8%) patients. A total of 642 (70.9%) patients began PORT within 6 weeks of radical surgery, with 436 (48.2%) having a TPT ≤85 days. The median RTT was 47 days (range, 39–56 days). A total of 556 (61.4%) and 905 (100%) patients completed RT within 7 and 8 weeks of surgery, respectively.

There were 116, 341, and 448 patients in the early missed RT, late missed RT, and RT as scheduled groups, respectively. The reasons for early missed RT were as follows: operational causes (machine breakdown, public holidays; n = 69), treatment plan modifications because of physician’s decision (including re-simulation and re-optimization of the treatment plan due to modifications of the facial profile, physician’s discretion, missing of fixation cast or oral bite block; n = 13), patients taking leaves (unknown reasons; n = 34), frailty (hospital admissions, weight loss; n = 18), and mixed reasons (operational causes and other reasons; n = 32).

The median duration of follow-up was 6.1 years (range, 0.2–13.9 years), during which 451 (49.8%) patients died. The 2-, 3-, and 5-year OS rates in the entire study cohort were 71.1, 67.6, and 60.6%, respectively. The 5-year OS rates according to pathologic stage (AJCC Eight Edition Staging Manual) are shown in Fig. 1a. Kaplan-Meier survival plots of OS, FFDM, and LRCR according to the patterns of missed RT sessions (early missed RT, late missed RT, RT as scheduled) are depicted in Fig. 1b−d. The 5-year OS rates in patients with early missed RT, late missed RT, and RT as scheduled were 53.0, 58.1, and 64.5%, respectively. The 5-year OS rate of patients with early missed RT was significantly lower than of patients with RT as schedule (p = 0.021). The 5-year FFDM rates in patients with early missed RT, late missed RT, and RT as scheduled were 76.1, 84.3, and 82.2%, respectively. Patients with early missed RT had a significantly higher rate of distant metastasis when compared with patients with late missed RT (p = 0.048). Finally, the 5-year LRC rates in patients with early missed RT, late missed RT, and RT as scheduled were 70.1, 69.1, and 77.8%, respectively. Patients with late missed RT had a significantly lower local-regional control rate when compared with patients with RT as scheduled (p = 0.011).

Table 2 summarizes the results of univariate analysis of variables associated with 5-year OS, FFDM, and LRCR rates. In addition to other variables, early missed RT (versus RT as scheduled) was identified as a significant adverse predictor of OS. Moreover, early missed RT (versus late missed RT) was significantly associated with a less favorable FFDM. After allowance for potential confounders in multivariate analysis (Table 3), early missed RT (versus RT as scheduled) retained its independent adverse prognostic significance for OS (HR = 1.201; 95% CI: 0.978–1.474; p = 0.008). Notably, early missed RT (versus RT as scheduled) was also independently associated with a less favorable FFDM (HR = 1.644; 95% CI: 1.047–2.583; p = 0.031).