Predictors of severe long-term toxicity after re-irradiation for head and neck cancer

Oral Oncology 60 (2016) 32–40

Contents lists available at ScienceDirect

Oral Oncology journal homepage: www.elsevier.com/locate/oraloncology

Predictors of severe long-term toxicity after re-irradiation for head and neck cancer Jae Y. Lee a,1, Krithika Suresh a,d,1, Rebecca Nguyen a, Eli Sapir a, Janell S. Dow a, George S. Arnould a, Francis P. Worden b, Matthew E. Spector c, Mark E. Prince c, Scott A. McLean c, Andrew G. Shuman c, Kelly M. Malloy c, Keith Casper c, Carol R. Bradford c, Matthew J. Schipper a,d, Avraham Eisbruch a,⇑ a

Department of Radiation Oncology, University of Michigan, Ann Arbor, MI, United States Department of Internal Medicine, Hematology/Oncology, University of Michigan, Ann Arbor, MI, United States Department of Otolaryngology-Head and Neck Surgery, University of Michigan, Ann Arbor, MI, United States d Department of Biostatistics, University of Michigan, Ann Arbor, MI, United States b c

a r t i c l e

i n f o

Article history: Received 19 May 2016 Received in revised form 23 June 2016 Accepted 24 June 2016

Keywords: Head and neck re-irradiation Interval to re-irradiation Re-irradiated volume Re-irradiation toxicity Dysphagia

a b s t r a c t Objective: To identify predictive factors of severe long-term toxicity after re-irradiation of recurrent/persistent or second-primary head and neck cancer. Methods: Outcomes and treatment plans of patients who underwent modern IMRT based re-irradiation to the head and neck from 2008–2015 were reviewed. Co-variables including demographic, clinical and oncologic factors, as well as interval to re-irradiation and re-irradiated planning tumor volume (PTV) were analyzed as predictors of developing severe (CTCAE grade P 3) long-term toxicity with death as a competing risk. Results: A total of 66 patients who met inclusion criteria were eligible for analysis. A median reirradiation dose of 70 Gy was delivered at a median of 37.5 months after initial radiotherapy. Reirradiation followed surgical resection in 25 (38%) patients, and concurrent chemotherapy was delivered to 41 (62%) patients. Median follow-up after re-irradiation was 23 months and median overall survival was 22 months (predicted 2 year overall survival 49%). Of the 60 patients who survived longer than 3 months after re-irradiation, 16 (25%) patients experienced severe long-term toxicity, with the majority (12 of 16) being feeding tube -dependent dysphagia. In multivariable analysis, shorter intervals to reirradiation (<20 months) and larger re-irradiated PTVs (>100 cm3) were independent predictors of developing severe long-term toxicity. Patients with longer disease-free intervals and smaller PTVs had a 94% probability of being free of severe toxicity at two years. Conclusion: Selection of patients with longer re-irradiation intervals and requiring smaller re-irradiated PTVs can independently predict avoidance of severe long-term toxicity. Ó 2016 Elsevier Ltd. All rights reserved.

Introduction Despite aggressive multimodality treatments involving a combination of surgery, radiation, and chemotherapy, a large subset of patients with head and neck squamous cell carcinoma (HNSCC) face the challenge of recurrent or second primary tumors that arise in tissues that have been previously irradiated. Locoregional recurrences account for approximately half of failures in patients treated with definitive chemoradiation, though control rates are much ⇑ Corresponding author at: 1500 East Medical Center Dr., Ann Arbor, MI 48109, United States. E-mail address: [email protected] (A. Eisbruch). 1 These authors contributed equally to this manuscript. http://dx.doi.org/10.1016/j.oraloncology.2016.06.017 1368-8375/Ó 2016 Elsevier Ltd. All rights reserved.

higher in human papillomavirus (HPV)-positive oropharyngeal cancers [1,2]. However, given the traditional non-viral risk factors associated with HNSCC, patients without recurrent cancers still have an 18% risk of developing a second primary tumor [3]. Ultimately, locoregional tumor progression is a major source of severe morbidity and is the predominant cause of death in patients with head and neck cancer [4–6]. The preferred management of recurrent or second primary cancers in previously irradiated tissue is surgical resection; however, a significant proportion of patients are not eligible for surgical resection due to myriad patient and oncological variables [7]. Even in the subset of operable patients, outcomes with salvage surgery alone remain poor and disease-free survival is improved with adjuvant re-irradiation with or without concurrent chemotherapy in

J.Y. Lee et al. / Oral Oncology 60 (2016) 32–40

highly selected cohorts [8]. As a result, a growing cadre of patients may be candidates for head and neck re-irradiation [9,10]. Beyond tumor control and survival, quality of life and survivorship issues are increasing in importance for patients with head and neck cancer [11]. Although survivorship issues are obviously relevant for patients with favorable prognosis, quality of life concerns are arguably equally or more important for patients with incurable advanced disease. Treatment-related toxicities associated with head and neck reirradiation can be substantial. However there is a dearth of data exploring how to predict their development, and thus to best select candidates who are likely to benefit. Patient and oncological factors are clearly overlapping and inter-related, but two disease-related factors that appear to be associated with outcomes of interest are the time interval to re-irradiation, and the recurrent tumor volume [12–14]. While we postulate that these factors can also impact toxicity, supportive data are scant. An important obstacle to reliably identify predictors of toxicity has been inconsistent and noncomprehensive tracking of treatment-related sequelae [15]. Since 2008, it has been our institutional policy to prospectively assess and grade head and neck radiation-related toxicities across a range of domains at each on-treatment visit. With this systematic practice in place, we maximized our ability to detect the true incidence of toxicity due to head and neck re-irradiation and identify independently predictive factors thereof. Patients and methods Inclusion criteria After institutional review board approval, we retrospectively reviewed the medical records, imaging, and treatment plans of all eligible patients who underwent radiation at our facility between 2008 and 2015. Inclusion criteria specified patients with biopsy-proven recurrent/persistent or second primary head and neck cancer (all histologies) treated with head and neck reirradiation with overlapping radiation fields, with a minimum biologically equivalent re-irradiation dose of 56 Gy. All re-irradiation plans were delivered via intensity modulated radiotherapy (IMRT) or volumetric modulated arc therapy (VMAT), and the majority was treated with curative intent. All tumors were restaged prior to re-irradiation with complete history and physical examination, tissue biopsy, and imaging with either computed tomography (CT), magnetic resonance imaging (MRI), and/or positron emission tomography (PET)-CT. All patients were evaluated in a multidisciplinary fashion and the clinical rationale for re-irradiation included patients who were not surgical candidates (due to unresectable disease, medical comorbidities or patient/surgeon discretion), or high risk features after surgery (positive margins, extracapsular nodal extension).

33

Concurrent chemotherapy with re-irradiation was delivered after multidisciplinary tumor board discussion, ultimately at the discretion of the treating medical oncologist. Regimens generally included weekly cisplatin, or carboplatin with or without paclitaxel. In selected cases, patients received concurrent capecitabine, 5-FU and carboplatin, cisplatin and etoposide, or cetuximab. Statistics The primary clinical endpoint was severe long-term toxicity, characterized as any long-term toxicity grade P 3 on the Common Terminology Criteria for Adverse Events (CTCAE) v4.0 scale that developed or persisted 90 days after completing re-irradiation. Freedom from severe long-term toxicity from end of reirradiation was estimated using the cumulative incidence function with death of any cause as a competing risk. Gray’s method was used for testing differences between these curves. Univariate and multivariable analyses were performed using Fine and Gray regression to identify association between severe long-term toxicity and the following prognostic factors: age at re-irradiation, sex, interval to re-irradiation, TNM staging of recurrence (AJCC 7th edition), overall stage of recurrence, salvage surgery prior to re-irradiation, concurrent chemotherapy with re-irradiation, dose at time of reirradiation, organ dysfunction at recurrence, and the volume of the re-irradiated PTV. The re-irradiated PTV was log-transformed in order to achieve normality when assessed as a continuous variable and a threshold of 100 cm3 was used when assessed as a dichotomized variable. Organ dysfunction included feeding tube dependence or soft tissue defect prior to re-irradiation as previously defined [12]. The interval to re-irradiation was considered as both continuous and dichotomized with a threshold of 20 months (the median interval to re-irradiation in patients with severe long-term toxicity). For factors with significant association, cumulative incidence functions were used to estimate time to severe long-term toxicity across factor levels. Secondary clinical endpoints of interest were overall survival, measured from last day of re-irradiation to date of death or censoring, locoregional recurrence (LRR)-free survival (measured from the last day of re-irradiation to date of relapse), death, and acute toxicity, which was prospectively graded on the CTCAE scale during on-treatment visits. Survival functions were estimated by the Kaplan–Meier method. The log-rank test was used to compare survival curves between groups. Univariate and multivariable survival analyses were performed with a Cox proportional hazards model to identify significant association between the survival rates and the prognostic factors described above in addition to Charlson comorbidity index [17], use of chemotherapy for initial disease, and recurrent disease sites. Model selection for multivariable analyses was conducted using forward stepwise selection. The statistical analyses were conducted using R 3.2.2 software. Statistical significance was determined with a of 0.05.

Treatment details Results In the case of gross disease, the re-irradiation volume was generally the gross target volume (GTV) with a 5 mm expansion directly to planning target volume (PTV) [16]. In cases with multifocal gross disease, a high risk clinical target volume (CTV) expansion of 5 mm was created around GTVs with a 3 mm margin for the PTV. Comprehensive elective nodal re-irradiation was generally avoided. In post-operative cases, the high risk CTV was defined using pre-operative and post-operative imaging with a 3 mm margin for the PTV. Planning objectives aimed to deliver the prescription dose to 95% of the PTV. Daily image guidance with cone beam CT and immobilization with a 5-point thermoplastic mask was used for all patients.

Patient characteristics A total of 66 patients underwent re-irradiation with demographic, clinical, and treatment characteristics summarized in Table 1. Re-irradiation was delivered at a median of 37.5 months after the initial course of radiotherapy with a median follow up of 23 months. There were 3 patients (5%) with new primaries in a previously irradiated field, and 20 patients (30%) were treated with re-irradiation for a 2nd or 3rd recurrent cancer. All but 4 patients were treated with curative intent with a median dose of 70 Gy and 4 patients received aggressive palliation with 50 Gy

34

J.Y. Lee et al. / Oral Oncology 60 (2016) 32–40

Table 1 Patient Characteristics.

Table 1 (continued) Characteristic

Characteristic

n or median

Percentage or range

Total patients

66

Patient characteristics Median age at re-irradiation

62.3

45.7–84.3

Gender Male Female

53 13

80.3 19.7

Charlson score at recurrence 0 1 2 3 4

42 18 3 2 1

63.6 27.3 4.5 3.0 1.5

Organ dysfunction at recurrence No Yes

36 30

54.5 45.5

Number of recurrences 0 1 2 3 Median interval to re-irradiation (months)

3 43 18 2 37.5

4.5 65.2 27.3 3.0 3–304

Tumor characteristics Recurrent disease sites Nasopharynx Oral cavity Oropharynx Larynx Hypopharynx Neck only Other

4 10 8 11 3 16 14

6.1 15.2 12.1 16.7 4.5 24.2 21.2

T stage at recurrence 0 1 2 3 4 Missing

15 5 13 6 18 9

22.7 7.6 19.7 9.1 27.3 13.6

N stage at recurrence 0 1 2 3 Missing

23 9 25 2 7

34.8 13.6 37.9 3.0 10.6

M stage at recurrence 0 1 Missing

57 3 6

86.4 4.5 9.1

Stage 1 2 3 4 Missing

2 6 10 43 5

3.0 9.1 15.2 65.2 7.6

Histology Squamous cell carcinoma Acinic cell carcinoma Esthesioneuroblastoma

61 3 2

92.4 4.5 3.0

Viral markers at recurrence HPVHPV+ Not tested

6 10 14

20.0 33.3 46.7

Treatment characteristics Salvage surgery prior to re-irradiation No Yes For a prior recurrence Median dose at RT2 (Gy)

28 25 13 70

42.4 37.9 19.7 50–74

Concurrent chemo with re-irradiation No Yes Carboplatin/Paclitaxel Cisplatin Carboplatin Cetuximab Cisplatin/Etoposide Carboplatin/5-FU Capecitabine Median re-irradiated planning target volume, PTV (cc)

n or median

Percentage or range

25 41 13 12 11 2 1 1 1 119.4

37.9 62.1 19.7 18.2 16.7 3.0 1.5 1.5 1.5 16.4–1141

delivered in 2.5 Gy fractions. Re-irradiation was delivered postoperatively for 25 patients (38%), and with concurrent chemotherapy for 41 patients (62%). Re-irradiation tolerance Acute toxicities were mild-moderate (grade 1–2) in 40 (60%) of 66 patients (Table 2A), with the majority of mild-moderate toxicities comprising mucositis and/or dermatitis. Severe acute toxicity (grade 3–5) was observed in 15 (23%) patients (6 with dysphagia, 4 with dermatitis, 4 with mucositis, and 1 death). Three of the six patients with acute grade 3 dysphagia persisted with long-term feeding tube dependence. There was one possible treatmentrelated death in a patient with recurrent laryngeal cancer who died shortly after completing antibiotics for aspiration pneumonia. No patients experienced an acute carotid blowout. For long-term toxicity, the 60 patients who survived longer than 3 months after reirradiation and adequate follow-up were analyzed (Table 2B). 16 (27%) patients suffered severe long-term toxicity, with the majority

Table 2 (A) Acute toxicities and (B) long-term toxicities. As some patients suffered from more than 1 toxicity, the total number of patients experiencing a particular grade is shown separately.

a

N = 66

Grade

Acute toxicity

1

2

3

4

5

A Fatigue Mucositis Dysphagia/Odynophagia Nausea Dermatitis Dysgeusia Xerostomia Worst overall Worst overall%

6 15 11 4 14 5 5 14 21.2

1 13 5 1 16 2 5 26 39.4

0 4 6 0 4 0 0 14 21.2

0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 1a 1.5

N = 60

Grade

Long-term toxicity

1

2

3

4

5

B Dental/Osteoradionecrosis Fibrosis Trismus Mucositis Dysphagia Soft tissue damage Dysgeusia Dysphonia Xerostomia Worst overall Worst overall%

0 1 0 0 1 0 2 0 3 2 3.3

1 11 13 4 6 0 3 3 13 29 48.3

1 0 2 0 12 3 0 0 1 15 25.0

0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 1 0 0 0 1 1.7

Aspiration pneumonia

J.Y. Lee et al. / Oral Oncology 60 (2016) 32–40

(12 of 16) experiencing feeding tube dependent dysphagia. By last follow-up, 21 (35%) of 60 patients required feeding tubes, which also includes patients with severe dysphagia prior to reirradiation. There were 3 additional long-term grade 3 toxicities independent of dysphagia: vertebral osteomyelitis and tracheitis within the treatment field requiring IV antibiotics and osteoradionecrosis of the jaw. One patient died due to carotid bleeding in the setting of tumor recurrence 3.5 months after re-irradiation. Predictors of severe long-term toxicity Both interval to re-irradiation and the size of the re-irradiated volume were able to independently predict severe long-term toxicity. At 2 years after re-irradiation, 80% of patients with long intervals to re-irradiation remained free of severe toxicity versus 47% of patients with shorter intervals (P = 0.02, Fig. 1A). Patients with smaller re-irradiated volumes (PTV 6 100 cm3) were less likely to experience severe long-term toxicity than patients with larger volumes (13% vs. 36%) at 2 years (P = 0.01, Fig. 1B). This size threshold was chosen to more closely approximate tumor sizes in patients treated with SBRT re-irradiation in other reports, although our threshold is larger than the sizes reported in these other studies [18,19]. Stratifying patients based on interval and PTV size, four different sub-groups were identified with varying risks of severe toxicity (Fig. 1C). Fine and Gray competing risk regression models for severe longterm toxicity are presented in Table 3. Univariate analyses identified shorter intervals to re-irradiation and larger re-irradiated volumes as significant predictors of severe long-term toxicity. Time interval to re-treatment was a significant predictor of severe

35

long-term toxicity when treated as both a continuous predictor (P = 0.03) and when dichotomized to interval 620 months vs >20 months (P = 0.02). In a multivariable analysis, shorter interval to re-irradiation and larger re-irradiated volume were significant predictors of developing severe late toxicity (Table 3). Patients with the highly favorable characteristics of a long interval to reirradiation and small PTVs are predicted to have a 95% probability of being alive and free of severe toxicity at 2 years, whereas patients with short intervals and large PTVs were predicted to have a 31% probability of being alive and free from severe toxicity (Fig. 2). Overall survival With a median follow-up time of 23 months, the median survival after completing re-irradiation was 22 months (95% CI: 12.8, not reached) for all patients. The 1- and 2-year survival estimates were 64% (95% CI: 53, 77) and 49% (95% CI: 38, 65), respectively. The median overall survival for patients with organ dysfunction at recurrence and those without was 6.7 months and 40.3 months, respectively (P < 0.001, Fig. 3). Classifying patients based on interval to re-irradiation and organ dysfunction, the median overall survival for patients with long intervals without organ dysfunction is 40.3 months compared with 5.4 months for patients with short intervals and organ dysfunction (P < 0.001, Fig. 3E). Cox proportional hazards models for overall survival are shown in Table 4. Univariate analyses identified increasing age at reirradiation and organ dysfunction at recurrence as significant factors associated with reduced overall survival. In the multivariable analysis, longer dichotomized retreatment interval was significantly

Fig. 1. Kaplan-Meier curves showing freedom from severe long-term toxicity based on (A) interval to re-irradiation and (B) PTV size, and (C) both.

36

J.Y. Lee et al. / Oral Oncology 60 (2016) 32–40

Table 3 Univariate and multivariable analysis for severe long-term toxicity. Variable

Age at re-irradiation Gender Male (vs. Female) Organ dysfunction at recurrence Number of recurrences Interval to re-irradiation, months (continuous) Interval to re-irradiation >20 mo. (vs. 620 mo.) Recurrent T stage Recurrent N stage Recurrent M stage Recurrent stage Salvage surgery prior to re-irradiation Re-irradiation dose Concurrent chemotherapy with re-irradiation log2 re-irradiated volume, cm3 (continuous) PTV P 100 cm3 (vs. PTV < 100 cm3)

Univariate analysis

Multivariable analysis

HR

95% CI

P

0.94 1.66 1.41 1.45 0.99 0.32 1.23 1.41 1.22 1.16 1.64 0.99 0.77 1.88 5.55

(0.89, 1.00) (0.46, 5.98) (0.57, 3.54) (0.71, 2.98) (0.98, 0.999) (0.13, 0.81) (0.76, 1.98) (0.76, 2.60) (0.16, 9.10) (0.66, 2.03) (0.57, 4.72) (0.91, 1.08) (0.29, 2.01) (1.41, 2.51) (1.43, 21.5)

.05 .44 .46 .31 .03 .02 .40 .28 .84 .61 .36 .85 .59 <.001 0.01

HR

95% CI

P

0.99 0.30

(0.98, 0.999) (0.11, 0.80)

.04a .02a

1.73

(1.36, 2.20)

<.001

a These estimates come from separate multivariable competing risks models that differ in the way that re-treatment interval was coded (i.e., continuous vs. dichomtomized). Estimates for other variables come from the multivariable analysis with dichotomized re-treatment interval and do not differ substantially in the analysis with continuous re-treatment interval.

is 31.1 months for patients with longer intervals without organ dysfunction compared to 2.7 months for patients with shorter intervals and organ dysfunction (Fig. 3F). The best fitting multivariable model identified shorter intervals to re-irradiation, salvage surgery prior to re-irradiation, and absence of organ dysfunction as independent predictors of LRRfree survival (Table 4). In a predictive multivariable model with organ dysfunction and dichotomized re-irradiation interval, patients with longer intervals to re-irradiation and absence of organ dysfunction have a 2-year LRR-free predicted survival of 64% compared to 2% for patients with shorter intervals and organ dysfunction (Fig. 4B).

Discussion Fig. 2. Predicted freedom from severe long-term toxicity based on interval to reirradiation and PTV size.

associated with improved survival, with a hazard ratio for death of 0.40 (95% CI: 0.19, 0.81) in patients who received re-treatment >20 months after initial radiation compared to patients with retreatment interval 620 months (Table 4). Organ dysfunction and increased age at re-irradiation were also significantly associated with death in the multivariable analysis. In a predictive multivariable model with organ dysfunction and dichotomized reirradiation interval, patients with both longer intervals to reirradiation as well as absence of organ dysfunction have a 2-year predicted overall survival of 77% compared with 8% for patients with shorter interval to re-irradiation and organ dysfunction (Fig. 4A). Locoregional recurrence (LRR)-free survival The median LRR-free survival after re-irradiation was 19.8 months (95% CI: 18.5, 40.0) for all patients. The 1- and 2year LRR-free survival rates were 83% (95% CI: 71, 97) and 42% (95% CI: 26, 66), respectively. Median LRR-free survival for patients with intervals >20 months was 15.7 months versus 3.2 months for patients with shorter intervals (P = 0.02, Fig. 3) In addition, median LRR-free survival for patients without organ dysfunction was 31.1 months versus 3.2 months for patients with organ dysfunction (P < 0.001, Fig. 3). Classifying patients based on interval to re-irradiation and organ dysfunction, the median LRR-free survival

Several prior studies have demonstrated that shorter intervals to head and neck re-irradiation are associated with poor survival, reflective of the fact that aggressive tumors frequently recur rapidly [12,13,20]. Consistent with this, we have identified shorter re-treatment intervals as an independent predictor of poor survival. Much effort has been made on refining the identification of patients who may derive the greatest survival benefit from head and neck re-irradiation [12,21]. As this selection process continues to improve, it has become clear that a subset of patients can achieve meaningful long-term survival. For these patients, very little is known regarding predictors of developing long-term toxicity. In an era in which patient-centered decision-making is the hallmark of cancer care, focusing on functional outcomes as well as oncologic metrics among our most vulnerable and high-risk is a fundamental responsibility. One challenge for assessing composite outcomes of interest in the past has been inconsistent and non-systematic tracking of toxicities associated with head and neck re-irradiation. By prospectively grading radiation-related toxicities among multiple domains (fatigue, weight loss, mucositis, esophagitis, nausea, dermatitis, dysgeusia, and xerostomia) at each on-treatment visit, we were able to benefit from comprehensive and longitudinal tracking and management of treatment related toxicities. With this practice, we have identified interval to re-irradiation as well as the volume of re-irradiated tissue as significant predictors of severe long-term toxicity. For patients previously treated with head and neck radiotherapy, optimal management of recurrent or second metachronous

J.Y. Lee et al. / Oral Oncology 60 (2016) 32–40

37

Fig. 3. Kaplan-Meier curves for overall survival and locoregional recurrence-free survival based on (A and B) interval to re-irradiation, (C and D) organ dysfunction and (E and F) both.

primary head and neck cancers is guided by few prospective reports and mainly based on clinical experience and numerous retrospective series in the literature. Although no randomized data exist, salvage surgery is generally considered first-line treatment of disease arising in an irradiated field when feasible. However, only a minority of patients present with resectable disease at recurrence and with careful selection, re-irradiation offers the only potentially curative option for some of these patients. Our experience with IMRT based re-irradiation for both unresectable and post-operative patients resulting in 2 year overall survival of 49% (locoregional recurrence-free survival 42%) with 27% of patients experiencing severe long-term toxicity compares favorably to that of benchmarks set in the literature, and informs how we frame oncologic as well as functional outcomes and expectations. The only prospective randomized trial evaluating re-irradiation with reported results comes from the Group d’Etude des Tumeurs

de la Tête et Cou (GETTEC) and Groupe d’Oncologie Radiothérapie Tête et Cou (GORTEC) which randomized patients who postoperatively were considered at high risk to either a split course of 60 Gy re-irradiation concurrently with 5-FU and hydroxyurea or initial observation [8]. The re-irradiated arm showed superior diseasefree survival (DFS) (HR 1.68) but not overall survival (approximately 45% at 2 years). Late toxicities in the re-irradiation arm was 39% at 2 years and 8% of patients died of treatment-related causes. It is important to note that all patients on that trial underwent R0 or R1 resections whereas only 38% of patients in our series were re-irradiated post-operatively. On the other end of the spectrum are trials examining the role of re-irradiation for unresectable patients. In RTOG 9610 patients were treated with another split course hyperfractionated 60 Gy delivered concurrently with hydroxyurea and 5-FU [20]. With this intensive treatment strategy, there was a 7.6% rate of treatment related deaths and a 56% rate of

38

J.Y. Lee et al. / Oral Oncology 60 (2016) 32–40

Table 4 Univariate and multivariable analysis of locoregional recurrence-free survival and overall survival. Variable

Locoregional recurrence-free survival Univariate analysis

Age at re-irradiation Gender Male (vs. Female) Charleson score Organ dysfunction at recurrence Number of recurrences Interval to re-irradiation, months (continuous) Interval to re-irradiation >20 mo. (vs. 620 mo.) Recurrent disease sites (vs. oral cavity) Nasopharynx Oropharynx Larynx Hypopharynx Neck only Other Recurrent T stage Recurrent N stage Recurrent M stage Recurrent stage Viral status (HPV+ vs. HPV ) Salvage surgery prior to re-irradiation Re-irradiation dose Concurrent chemotherapy with re-irradiation log2 re-irradiated volume, cm3 (continuous)

HR

95% CI

P

1.03 0.82 0.99 3.05 1.14 0.997

(0.99, 1.06) (0.38, 1.78) (0.70, 1.42) (1.65, 5.66) (0.71, 1.82) (0.993, 1.00)

.12 .61 .97 <.001 .60 .14

0.47

(0.26, 0.87)

.02

0.45 0.44 0.50 0.28 0.66 0.88 1.03 1.32 1.43 1.55 1.04 0.54 0.96 1.08

(0.07, 2.73) (0.11, 1.74) (0.14, 1.76) (0.03, 2.45) (0.22, 1.99) (0.30, 2.62) (0.83, 1.27) (0.93, 1.87) (0.34, 5.97) (0.96, 2.50) (0.25, 4.36) (0.28, 1.06) (0.91, 1.01) (0.58, 2.04)

.39 .24 .28 .25 .46 .82 .80 .12 .63 .08 .96 .07 .14 .80

1.09

(0.83, 1.41)

.54

Overall survival

Multivariable analysis

Univariate analysis

HR

HR

95% CI

P

HR

95% CI

P

1.05 0.97 1.14 4.41 1.15 0.999

(1.02, 1.09) (0.42, 2.23) (0.77, 1.68) (2.15, 9.02) (0.71, 1.87) (0.995, 1.00)

.004 .94 .52 <.001 .57 .78

1.06

(1.02, 1.11)

.006

4.53

(2.16, 9.52)

<.001

0.58

(0.30, 1.12)

.11

0.40

(0.19, 0.81)

.01

0.45 0.44 0.50 0.28 0.66 0.88 1.06 1.17 1.27 1.71 1.14 0.67 0.95 0.80

(0.07, 2.73) (0.11, 1.74) (0.14, 1.76) (0.03, 2.45) (0.22, 1.99) (0.30, 2.62) (0.85, 1.33) (0.81, 1.70) (0.30, 5.39) (0.93, 3.13) (0.21, 6.32) (0.33, 1.36) (0.90, 1.01) (0.41, 1.56)

.40 .24 .28 .25 .46 .82 .61 .40 .75 .08 .88 .26 .10 .51

1.16

(0.86, 1.55)

.33

95% CI

P

4.11

(2.12, 7.97)

<.001

0.47

(0.25, 0.89)

.02

0.45

(0.22, 0.92)

.03

Multivariable analysis

Fig. 4. Predicted (A) overall survival and (B) locoregional recurrence-free survival based on interval to re-irradiation and organ dysfunction.

acute grade 3–4 toxicities. Late toxicities were reported in 22% of patients, though 70% of patients had feeding tubes at last follow up, perhaps reflecting both the poor baseline function of unresectable patients and a highly intensive re-irradiation regimen. Overall survival was 15% at 2 years and 73% of deaths were cancer related. Re-irradiation intervals longer than 1 year were associated with improved survival (median 5.8 months vs. 9.8 months). RTOG 9911 used a similar schedule for re-irradiation with cisplatin/paclitaxel instead and granulocyte colony-stimulating factor during the rest weeks [22]. A similar rate of treatment related deaths (8%) was seen and grade 3–4 toxicities were noted in 34% of patients. 2 year overall survival was 26% and was improved compared with that of RTOG 9610. Our results are consistent with other modern retrospective analyses. Similar to Tanvetyanon et al., we also identified organ dysfunction (feeding tube dependence, tracheostomy, soft tissue defect) prior to re-irradiation as a marker of poor survival [12].

Such patients are ill-equipped to fully maximize their nutritional and performance status and experience poor treatment tolerance or death from non-cancer related causes. However, our analyses did not find a correlation between comorbidity and survival. While the frequency of significant comorbidity (Charlson index P 1) was similar in both studies (both 36%), the rates of severe long-term toxicities were lower in our experience (27% vs. 48%). It is possible that the contribution of comorbidities to survival is amplified in the context of severe treatment-related toxicity. Furthermore, we also identified increasing age as an independent predictor of overall survival, which also likely captures unmeasured comorbidity information. In a large retrospective analysis, Choe et al. identified absence of prior exposure to chemoradiation and long reirradiation intervals as favorable prognostic factors, among others [13]. In our series, patients who were not treated previously with concurrent chemoradiation harbored either early glottic larynx cancer or received adjuvant radiotherapy for microscopic disease.

J.Y. Lee et al. / Oral Oncology 60 (2016) 32–40

Both features were felt to be significant confounders for survival, and we did not include previous treatment with chemoradiation as a variable. We also did not observe a dose effect on survival in our results, likely because we met or exceeded a commonly accepted cutoff of 60 Gy in virtually all patients. Re-irradiation toxicity profiles also appear promising from more modern series using hypo-fractionated stereotactic body radiotherapy (SBRT), particularly those utilizing an every other day treatment schedule [23]. In a report of the Pittsburgh experience using SBRT with or without concurrent cetuximab, cumulative acute and late toxicity rates (grade 3 or higher) were 12% and 7%, respectively [19]. However, patients amenable to reirradiation via SBRT typically harbor smaller tumor volumes (median 31 cm3) and follow up has typically been much shorter (median 6 months) [19]. Re-irradiation with SBRT by the Pittsburgh group typically involved no expansion from gross tumor to PTV, though 3–5 mm expansions for PTV are now permitted in their current protocols [24,25]. When patients with smaller re-irradiated volumes were examined separately in our cohort, the rate of severe long-term toxicity was comparably low at 13%, although our size threshold for small volumes was <100 cm3 with a median follow up of 23 months. In the Pittsburgh experience, tumor volume was also identified as a predictor for acute toxicity but not longterm toxicity [19]. It is possible that the role of tumor size to long-term toxicity might be better interrogated with a wider distribution of tumor sizes. Consistent with this, Georgetown has also published their experience with SBRT for head and neck reirradiation, initially with an 11% rate of grade 4–5 (grade 3 and higher not reported) long-term toxicity with a median treated volume of 75 cm3 [18]. A subsequent report with additional patients to the original cohort showed an improved 5.9% rate of grade 3 and higher toxicity, though this was in the context of a short median survival of 8.6 months and 40% of patients being treated with palliative intent only [26]. The limitations of our study relate mainly to its retrospective design. There are attendant caveats regarding selection bias, although clinical judgment regarding eligibility for highly morbid treatment of advanced disease is obligatory in this patient population, and will undoubtedly inform decision-making. Although acute toxicities were prospectively graded during on-treatment visits, late toxicities were retrospectively graded. However, all severe late toxicities reported here were unambiguously severe and objectively determined; we anticipate that prospective grading would not be discordant. It still remains possible that the incidence of late toxicity is underreported, though our lack of many patients who were lost to follow-up along with our prospective grading of acute toxicity and its resultant attention at follow up minimizes this possibility. Despite a relatively low patient number, the fact that re-irradiation interval and re-irradiated volume were identified as predictors for severe long term toxicity is compelling and will require validation in a larger (preferably prospective) cohort in order to fundamentally inform treatment guidelines.

Conclusion In addition to re-irradiated volume, to our knowledge we have identified for the first time that the interval to re-irradiation is also an independent predictor of severe long-term toxicity after reirradiation for head and neck cancer. Careful selection of patients by age, organ dysfunction, and re-irradiation interval can result in meaningful long-term survival in a subset of patients. As we collectively strive to better select which patients are most likely to benefit from head and neck re-irradiation, consideration of these variables can assist clinicians in mitigating treatment-related toxicities.

39

Conflict of interest statement All authors declare no conflicts.

References [1] Blanchard P, Baujat B, Holostenco V, Bourredjem A, Baey C, Bourhis J, et al. Meta-analysis of chemotherapy in head and neck cancer (MACH-NC): a comprehensive analysis by tumour site. Radiother Oncol J Eur Soc Ther Radiol Oncol 2011;100:33–40. http://dx.doi.org/10.1016/j.radonc.2011.05.036. [2] Ang KK, Harris J, Wheeler R, Weber R, Rosenthal DI, Nguyen-Tân PF, et al. Human papillomavirus and survival of patients with oropharyngeal cancer. N Engl J Med 2010;363:24–35. http://dx.doi.org/10.1056/NEJMoa0912217. [3] Cooper JS, Pajak TF, Rubin P, Tupchong L, Brady LW, Leibel SA, et al. Second malignancies in patients who have head and neck cancer: incidence, effect on survival and implications based on the RTOG experience. Int J Radiat Oncol Biol Phys 1989;17:449–56. [4] Hong WK, Bromer RH, Amato DA, Shapshay S, Vincent M, Vaughan C, et al. Patterns of relapse in locally advanced head and neck cancer patients who achieved complete remission after combined modality therapy. Cancer 1985;56:1242–5. [5] Kotwall C, Sako K, Razack MS, Rao U, Bakamjian V, Shedd DP. Metastatic patterns in squamous cell cancer of the head and neck. Am J Surg 1987;154:439–42. [6] Sciubba JJ. End of life considerations in the head and neck cancer patient. Oral Oncol 2009;45:431–4. http://dx.doi.org/10.1016/j.oraloncology.2008.06.001. [7] Mabanta SR, Mendenhall WM, Stringer SP, Cassisi NJ. Salvage treatment for neck recurrence after irradiation alone for head and neck squamous cell carcinoma with clinically positive neck nodes. Head Neck 1999;21:591–4. [8] Janot F, de Raucourt D, Benhamou E, Ferron C, Dolivet G, Bensadoun R-J, et al. Randomized trial of postoperative reirradiation combined with chemotherapy after salvage surgery compared with salvage surgery alone in head and neck carcinoma. J Clin Oncol 2008;26:5518–23. http://dx.doi.org/10.1200/ JCO.2007.15.0102. [9] Sher DJ, Haddad RI, Norris CM, Posner MR, Wirth LJ, Goguen LA, et al. Efficacy and toxicity of reirradiation using intensity-modulated radiotherapy for recurrent or second primary head and neck cancer. Cancer 2010;116:4761–8. http://dx.doi.org/10.1002/cncr.25305. [10] Nieder C, Andratschke NH, Grosu AL. Increasing frequency of reirradiation studies in radiation oncology: systematic review of highly cited articles. Am J Cancer Res 2013;3:152–8. [11] Ringash J. Survivorship and quality of life in head and neck cancer. J Clin Oncol 2015;33:3322–7. http://dx.doi.org/10.1200/JCO.2015.61.4115. [12] Tanvetyanon T, Padhya T, McCaffrey J, Zhu W, Boulware D, DeConti R, et al. Prognostic factors for survival after salvage reirradiation of head and neck cancer. J Clin Oncol 2009;27:1983–91. http://dx.doi.org/10.1200/ JCO.2008.20.0691. [13] Choe KS, Haraf DJ, Solanki A, Cohen EEW, Seiwert TY, Stenson KM, et al. Prior chemoradiotherapy adversely impacts outcomes of recurrent and second primary head and neck cancer treated with concurrent chemotherapy and reirradiation. Cancer 2011;117:4671–8. http://dx.doi.org/10.1002/cncr.26084. [14] Rwigema J-CM, Heron DE, Ferris RL, Andrade RS, Gibson MK, Yang Y, et al. The impact of tumor volume and radiotherapy dose on outcome in previously irradiated recurrent squamous cell carcinoma of the head and neck treated with stereotactic body radiation therapy. Am J Clin Oncol 2011;34:372–9. http://dx.doi.org/10.1097/COC.0b013e3181e84dc. [15] Ringash J, Bernstein LJ, Cella D, Logemann J, Movsas B, Murphy B, et al. Outcomes toolbox for head and neck cancer research: outcomes toolbox for head and neck cancer. Head Neck 2015;37:425–39. http://dx.doi.org/10.1002/ hed.23561. [16] Popovtzer A, Gluck I, Chepeha DB, Teknos TN, Moyer JS, Prince ME, et al. The pattern of failure after reirradiation of recurrent squamous cell head and neck cancer: implications for defining the targets. Int J Radiat Oncol 2009;74:1342–7. http://dx.doi.org/10.1016/j.ijrobp.2008.10.042. [17] Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373–83. [18] Unger KR, Lominska CE, Deeken JF, Davidson BJ, Newkirk KA, Gagnon GJ, et al. Fractionated stereotactic radiosurgery for reirradiation of head-and-neck cancer. Int J Radiat Oncol 2010;77:1411–9. http://dx.doi.org/10.1016/j. ijrobp.2009.06.070. [19] Vargo JA, Heron DE, Ferris RL, Rwigema J-CM, Kalash R, Wegner RE, et al. Examining tumor control and toxicity after stereotactic body radiotherapy in locally recurrent previously irradiated head and neck cancers: Implications of treatment duration and tumor volume: treatment duration in reirradiation head and neck cancers. Head Neck 2014. http://dx.doi.org/10.1002/hed.2346. [20] Spencer SA, Harris J, Wheeler RH, Machtay M, Schultz C, Spanos W, et al. Final report of RTOG 9610, a multi-institutional trial of reirradiation and chemotherapy for unresectable recurrent squamous cell carcinoma of the head and neck. Head Neck 2008;30:281–8. http://dx.doi.org/10.1002/ hed.20697. [21] Riaz N, Hong JC, Sherman EJ, Morris L, Fury M, Ganly I, et al. A nomogram to predict loco-regional control after re-irradiation for head and neck cancer.

40

J.Y. Lee et al. / Oral Oncology 60 (2016) 32–40

Radiother Oncol 2014;111:382–7. http://dx.doi.org/10.1016/j.radonc.2014. 06.003. [22] Langer CJ, Harris J, Horwitz EM, Nicolaou N, Kies M, Curran W, et al. Phase II study of low-dose paclitaxel and cisplatin in combination with split-course concomitant twice-daily reirradiation in recurrent squamous cell carcinoma of the head and neck: results of radiation therapy oncology group protocol 9911. J Clin Oncol 2007;25:4800–5. http://dx.doi.org/10.1200/JCO.2006.07.9194. [23] Yazici G, Sanlı TY, Cengiz M, Yuce D, Gultekin M, Hurmuz P, et al. A simple strategy to decrease fatal carotid blowout syndrome after stereotactic body reirradiaton for recurrent head and neck cancers. Radiat Oncol 2013;8:1. [24] Wang K, Heron DE, Clump DA, Flickinger JC, Kubicek GJ, Rwigema J-CM, et al. Target delineation in stereotactic body radiation therapy for recurrent head and neck cancer: a retrospective analysis of the impact of margins and

automated PET-CT segmentation. Radiother Oncol 2013;106:90–5. http://dx. doi.org/10.1016/j.radonc.2012.11.008. [25] Vargo JA, Ferris RL, Ohr J, Clump DA, Davis KS, Duvvuri U, et al. A prospective phase 2 trial of reirradiation with stereotactic body radiation therapy plus cetuximab in patients with previously irradiated recurrent squamous cell carcinoma of the head and neck. Int J Radiat Oncol 2015;91:480–8. http://dx. doi.org/10.1016/j.ijrobp.2014.11.023. [26] Kress M-AS, Sen N, Unger KR, Lominska CE, Deeken JF, Davidson BJ, et al. Safety and efficacy of hypofractionated stereotactic body reirradiation in head and neck cancer: long-term follow-up of a large series: Stereotactic reirradiation in head and neck cancer. Head Neck 2015;37:1403–9. http://dx.doi.org/10.1002/ hed.2376.