Impact of cisplatin dose intensity on human papillomavirus-related and -unrelated locally advanced head and neck squamous cell carcinoma

European Journal of Cancer 67 (2016) 174e182

Available online at www.sciencedirect.com

ScienceDirect journal homepage: www.ejcancer.com

Original Research

Impact of cisplatin dose intensity on human papillomavirus-related and -unrelated locally advanced head and neck squamous cell carcinoma Anna Spreafico a, Shao Hui Huang b, Wei Xu c, Roberta Granata d, Chen-Shin Liu b, John N. Waldron b, Eric Chen a, Jolie Ringash b, Andrew Bayley b, Kelvin K.W. Chan a, Andrew J. Hope b, John Cho b, Albiruni A.R. Razak a, Aaron Hansen a, Raymond Jang a, Bayardo Perez-Ordonez e, Ilan Weinreb e, Paolo Bossi d, Ester Orlandi d, Lisa F. Licitra d, Yuyao Song c, Brian O’Sullivan b, Lillian L. Siu a, John Kim b,* a

Division of Medical Oncology and Hematology, Princess Margaret Cancer Centre, University of Toronto, Canada Department of Radiation Oncology, Princess Margaret Cancer Centre, University of Toronto, Canada c Department of Biostatistics, Princess Margaret Cancer Centre, University of Toronto, Canada d Head and Neck Medical Oncology Department, Fondazione IRCCS, Istituto Nazionale dei Tumori, Milano, Italy e Laboratory Medicine Program, Princess Margaret Cancer Centre, University of Toronto, Canada b

Received 31 May 2016; received in revised form 16 August 2016; accepted 21 August 2016

KEYWORDS Human papillomavirus; Head and neck cancer; Survival; Chemoradiotherapy; Cisplatin

Abstract Aim: The aim is to evaluate the impact of cisplatin dose modification on outcomes of human papillomavirus (HPV)-related (HPVþ) and HPV-unrelated (HPV) locally advanced head and neck cancer (LAHNC) treated with chemoradiotherapy (CRT). Patients and methods: A pooled analysis was conducted of stage III/IV oropharyngeal cancer (OPC), carcinoma of unknown primary (CUP) and laryngo-hypopharyngeal cancer (LHC) patients treated with single-agent cisplatin CRT in 2000e2012 from two tertiary academic cancer centres. HPV status was determined by p16 staining and/or in situ hybridisation. LHC was assumed to be HPV. Unknown HPV status OPC/CUPs were excluded. Overall survival (OS) was calculated. Multivariable analysis (MVA) evaluated the impact of cisplatin dose intensity on survival for HPVþ and HPV cohorts separately.

* Corresponding author: Department of Radiation Oncology, Princess Margaret Cancer Centre, University Health Network, 610 University Avenue, Toronto, M5G 2M9, Ontario, Canada. Fax: þ1 416 946 6561. E-mail address: [email protected] (J. Kim). http://dx.doi.org/10.1016/j.ejca.2016.08.013 0959-8049/ª 2016 Elsevier Ltd. All rights reserved.

A. Spreafico et al. / European Journal of Cancer 67 (2016) 174e182

175

Results: A total of 404 HPVþ and 255 HPV LAHNC (481 OPC, 18 CUP, 160 LHC) patients were included. Median follow-up was 4.3 (0.5e11.9) years. Three-year OS for cisplatin <200, Z200, and >200 mg/m2 subgroups were 52%, 60%, and 72% (P Z 0.001) for the HPV and 91%, 90%, and 91% (P Z 0.30) for the HPVþ patients. MVA confirmed a survival benefit with cisplatin >200 mg/m2 for the HPV (hazard ratio [HR] 0.5, 95% confidence interval [CI]: 0.3e0.7, P < 0.001) but not for HPVþ (HR 0.6, 95% CI: 0.4e1.1, P Z 0.104). There was a superior OS trend in the HPVþ T4 or N3 high-risk subset (N Z 107) with cisplatin >200 mg/ m2 (HR 0.5, 95% CI: 0.2e1.1, P Z 0.07). Conclusions: A survival benefit of cisplatin dose >200 mg/m2 is evident for HPV LAHNC patients, but not for HPVþ cohort overall, although the T4 or N3 subset may benefit from a higher cumulative cisplatin dose. ª 2016 Elsevier Ltd. All rights reserved.

1. Introduction Chemoradiotherapy (CRT) is a standard treatment approach for locally advanced head and neck cancer (LAHNC). The most common regimen is three cycles of high-dose (100 mg/m2) cisplatin delivered concurrently with a 7-week radiotherapy (RT) course on days 1, 22 and 43. Weekly cisplatin (typically 40 mg/m2 per week) is an alternative schedule. The cisplatin dose intensity (the amount of drug delivered per unit time [1]) used in current clinical practice for LAHNC was derived empirically. LAHNC patients treated with CRT often experience moderate-to-severe side-effects limiting their tolerance to receive the intended cisplatin dose intensity. Chemotherapy modifications (dose reductions/delays/omissions) are common (23e40%) [2,3]). The consequence of cumulative cisplatin dose reduction is uncertain although some reports suggest a possible detrimental impact on survival [2,4]. Understanding the optimal cisplatin dose intensity is further complicated by the emergence of human papillomavirus-related (HPVþ) HNC. Although most HPVþ patients present with LAHNC, their outcomes are favorable compared to traditional smoking-related HPV-unrelated (HPV) patients with the same tumourenodeemetastasis (TNM) stage [5,6]. A large single institutional report [3] recently identified subgroups of HPVþ patients with excellent outcomes with RT alone. A reasonable corollary is that full-dose cisplatin may not be necessary for every HPVþ oropharyngeal cancer (OPC) patient. Treatment deintensification targeting low-risk HPVþ OPC and intensification for high-risk HPVþ OPC and HPV LAHNC are now under investigation. Understanding the effects of cisplatin dose intensity in standard clinical practice for HPVþ and HPV LAHNC separately remains very relevant. Furthermore, it is important to appreciate the needs of HPV patients who do not have the favorable prognosis of the numerically dominant HPVþ patients. Oncologic treatment principles may not apply equally to both groups. Hence, we conducted this pooled analysis of LAHNC patients treated with singleagent cisplatin CRT at the Princess Margaret Cancer

Centre (PMH) (Toronto, Canada) and Istituto Nazionale dei Tumori (INT) (Milan, Italy) to evaluate the impact of cisplatin dose intensity for HPVþ and HPV LAHNC patients and to identify patients who benefit most from high-dose cisplatin. 2. Methods 2.1. Study population After research ethics board approval, a pooled analysis was conducted of all newly diagnosed stage III/IV OPC, laryngo-hypopharyngeal cancer (LHC) and carcinoma of unknown primary (CUP) treated with definitive single-agent cisplatin CRT between 2000 and 2012. HPV status was determined for all OPC and CUP using either in situ hybridisation of HPV DNA or p16 immunostaining as a surrogate, following accepted criteria [7,8]. OPC and CUP patients with unknown HPV status or patients who received induction chemotherapy or other systemic agents were excluded. LHC was not routinely tested for p16 [9], but assumed to be negative as in landmark reports [2,5]. PMH patients were identified from a prospective HNC database [10]. INT patients were identified from an institutional database. All patients were staged according to the seventh edition UICC/AJCC TNM classification. 2.2. Treatment All patients were treated with intensity-modulated radiotherapy (IMRT) or three-dimensional conformal radiotherapy (3DRT) to a gross tumour dose of 66e70 Gy in 33e35 fractions over 6.5e7 weeks (2 Gy per fraction) (Table 1). IMRT was standard from 2005 to 2006 at PMH and INT, respectively. Concurrent cisplatin regimens were 1) high dose (100 mg/m2 three weekly on RT days 1, 22 and 43) or 2) weekly (40 mg/m2 for 7 weeks). The choice of a three weekly versus weekly cisplatin schedule was based on institutional guidelines taking into account patient factors including Zubrod Performance Scale (Zubrod PS) and comorbidities.

176

A. Spreafico et al. / European Journal of Cancer 67 (2016) 174e182

2.3. Follow-up and outcome assessment Local and regional recurrences were confirmed histologically, while distant metastasis (DM) was diagnosed by characteristic radiologic findings  histologic confirmation. At PMH, vital status was further linked to the Ontario Population-Based Cancer Registry.

2.4. Statistical analysis Comparisons of clinical characteristics used Fisher’s exact test and Pearson’s chi-square test for categorical variables and Student’s t-test for continuous variables. Overall survival (OS) (including death from any cause) and percutaneous endoscopic gastrostomy (PEG) tube dependency rates were estimated using KaplaneMeier methods. Locoregional control (LRC), distant control (DC), and grade 3e4 late toxicity according to Radiation Therapy Oncology Group/Eastern Cooperative Oncology Group criteria were calculated by the competing risk method (considering death without an event as a competing risk). Clinical end-points were compared by log-rank test between cisplatin <200, Z200 and >200 mg/m2 for HPVþ and HPV diseases, separately. All time-toevent end-points were calculated from date of diagnosis (histologic confirmation), except late toxicity and PEG dependency, which were calculated from date of CRT completion. Multivariable analysis (MVA) with Cox regression was performed on HPVþ and HPV separately to identify survival predictors, including the following variables: cisplatin dose (>200 versus 200 mg/m2), age (continuous), smoking pack-years, T and N categories, and RT technique (IMRT versus 3DRT). For the HPV cohort, we also included disease site (OPC/CUP versus LHC) as a variable. HPVþ and HPV cohorts were analysed using different dichotomisation of T (T4 versus T0eT3 for HPVþ; T3eT4 versus T0eT2 for HPV) and N (N2c versus N0eN2b for HPVþ; N2b-N3 versus N0eN2a for HPV) categories because of the reported differential prognostication of T and N categories for both diseases [11,12].

MVA subset analysis explored the impact of cisplatin dose intensity on HPVþ OPC patients grouped by one risk classification schema from Ang et al. [5] and another from O’Sullivan et al. [3]. Additional power calculations (PASS software package; NCSS, Kaysville, Utah) were conducted to estimate sample sizes required to reach statistical significance for the different risk subsets. All tests were two-sided and results were considered significant if the P-value was <0.05. 3. Results 3.1. Clinical characteristics A total of 659 (404 HPVþ and 255 HPV) of 1070 consecutive LAHNC patients (PMH Z 810 and INT Z 260) were eligible, of whom 60 (9%) were enrolled in single-agent cisplatin clinical trials (Fig. 1). OS was similar between PMH and INT for HPVþ (3year OS: 90% [87e94] versus 94% [87e100], p Z 0.20) and HPV (3-year OS: 61% [55e68] versus 76% [60e97], p Z 0.08) patients. The clinical characteristics are shown in Table 2. Twelve (3%) HPVþ and 19 (8%) HPV patients received weekly cisplatin, and there was no survival difference between weekly versus three weekly cisplatin (HPVþ: p Z 0.76; HPV: p Z 0.96). A similar proportion of HPVþ versus HPV patients received cumulative cisplatin 200 mg/m2 (56% versus 62%, P Z 0.14). Median cisplatin dose for HPVþ and HPV patients were 221 and 215 mg/m2, respectively. The main reasons for cisplatin dose reduction were declining Zubrod PS and myelotoxicity (Table 3) for both HPVþ and HPV patients. Median follow-up was 4.4 years (range 0.6e11.9) and 3.9 years (range 0.5e10.7) in HPVþ and HPV cohorts, respectively. 3.2. Impact of cisplatin dose intensity in 255 HPV patients Three-year OS for cisplatin <200 and Z200 mg/m2 subsets were similar (52% versus 60%, P Z 0.32) and both were much lower compared to the >200 mg/m2

Table 1 Typical dose targets for different RT techniques. Data source

RT dose levels

3DRT with sequential phase reducing field technique

IMRT with simultaneous integrated boost technique

PMH

     

     

     

INT

Gross target Intermediate dosea Microscopic dose Gross target Intermediate dosea Microscopic dose

70 60 50 70 60 50

Gy/35 Gy/30 Gy/25 Gy/35 Gy/30 Gy/25

fractions/7 fractions/6 fractions/5 fractions/7 fractions/6 fractions/5

weeks weeks weeks weeks weeks weeks

70 Gy/35 fractions/7 weeks 63 Gy/35 fractions/7 weeks 56 Gy/35 fractions/7 weeks 69.96 Gy/33 fractions/6.5 weeks 59.4 Gy/33 fractions/6.5 weeks 56.1 Gy/33 fractions/6.5 weeks

Abbreviations: PMH: Princess Margaret Cancer Centre cohort; INT: Istituto Nazionale dei Tumori cohort; RT: radiotherapy; 3DRT: threedimensional conformal radiotherapy; IMRT: intensity-modulated radiotherapy. a Intermediate dose target are ordinarily introduced to the treatment of small volume lymph node objects, especially when radiologically suspicious for involvement by tumour.

A. Spreafico et al. / European Journal of Cancer 67 (2016) 174e182

PMH Cohort

177

INT Cohort

Stage III-IV OPC, LHC, CUP N=810

Stage III-IV OPC, LHC, CUP N=260 Excluded 80 HPV status unknown • OPC (N=72) • CUP (N=8)

Excluded 154 HPV status unknown •OPC (N=98) •CUP (N=56)

Tumour HPV Ascertained N=656

Tumour HPV Ascertained N=180

Excluded 105 cases received other systemic agents

Excluded 72 cases received other systemic agent

Final Eligible Cases N=75 • HPV+: 51 • HPV−( tested): 18 • HPV− assumed: 6

Final Eligible Cases N=584 • HPV+: 353 • HPV− (tested): 109 • HPV− assumed: 122

Combined Eligible Cohorts N=659 HPV+: 404 HPV−: 255

Fig. 1. Diagram of selection process for study cohorts.

subset (72%) (Fig. 2A); therefore, we dichotomised HPV patients into cisplatin 200 mg/m2 (N Z 158, 62%) versus >200 mg/m2 (N Z 97, 38%) subgroups for further analysis. Apart from age, baseline characteristics between the two subgroups were similar (Table 2). In the MVA, cumulative cisplatin dose >200 mg/m2 significantly reduced mortality risk compared to those who received 200 mg/m2, hazard ratio (HR) 0.5 (95% confidence interval [CI]: 0.3e0.7, P < 0.001). IMRT reduced mortality risk compared to 3DRT (HR 0.6 [0.4e0.9], P Z 0.009) (Table 4). 3.3. Impact of cisplatin dose intensity in 404 HPVþ patients Three-year OS for cisplatin <200, Z200, and >200 mg/ m2 subsets were similar (93% versus 90%, versus 90%, P Z 0.76) (Fig. 2B). To be consistent with the HPV analysis, HPVþ patients were dichotomised into cisplatin 200 mg/m2 (N Z 227, 56%) versus >200 mg/ m2 (N Z 177, 44%) subgroups for further analysis. Unlike HPV patients, there was no significant difference between HPVþ 200 and >200 mg/m2 subgroups for 3-year OS (P Z 0.30), DFS (P Z 0.13), LRC (P Z 0.22), or DC (P Z 0.52) (Table 2). In MVA, the risk of death did not differ between cumulative cisplatin dose >200 versus 200 mg/m2 (HR

0.6 [0.4e1.1], P Z 0.104). Similar to HPV patients, IMRT reduced mortality risk (HR 0.4 [0.2e0.8], P Z 0.007) compared to 3DRT (Table 4). The result remained consistent when confined to the IMRT subset for both HPVþ and HPV diseases (Table 4). The additional exploratory MVA for the HPVþ OPC high mortality risk group (N2b-N3 and >10 smoking pack-year) defined by Ang et al. [5] demonstrated no survival difference between cisplatin >200 (N Z 82) and 200 mg/m2 (N Z 99) (3-year: 90% [83e97] versus 90% [84e97], P Z 0.76) (Fig. 2C) with an HR of 0.92 (95% CI: 0.41e2.07) (P Z 0.85). In contrast, a non-significant lower mortality risk for cisplatin >200 mg/m2 (N Z 53) and 200 mg/m2 (N Z 54) was found for the high DM risk subset (T4 or N3), defined by O’Sullivan et al. [3] (3year OS: 76% [64e89] versus 79% [68e93], P Z 0.15) (Fig. 2D) with the adjusted HR of 0.5 (95% CI: 0.2e1.1) by MVA (P Z 0.07). The power calculations for HPVþ risk groups indicated that if an additional 41 cases were accrued to the T4 or N3 high risk of DM [3] and death [11,12] subset, the OS difference for cisplatin >200 versus 200 mg/m2 would become significant (P < 0.05). For the T0-3N2c intermediate-risk subset, an additional 1567 cases would be required to reach significance. For the T03N0eN2b low-risk subset, equivalence testing showed no significant effect by cisplatin dose intensity

178

Table 2 Clinical characteristics and outcomes of 404 HPVþ and 255 HPV patients. Clinical characteristics

HPV, N (%)

Total

Cisplatin 200 mg/m

Cisplatin >200 mg/m

404 221 4.4 (0.6e11.9)

227 (56) 181 4.4 (0.8e11.5)

57 (27e81)

P-value

Total

Cisplatin 200 mg/m2

Cisplatin >200 mg/m2

P-value

177 (44) 283 4.5 (0.6e11.9)

0.57

255 215 3.9 (0.5e10.7)

158 (62) 174 3.8 (0.9e10.3)

97 (38) 280 4.4 (0.5e10.7)

0.46

57 (27e81)

55 (38e74)

0.012

60 (32e77)

60 (33e77)

58 (32e72)

0.002

89 (22) 315 (78)

48 (21) 179 (79)

41 (23) 136 (77)

0.63

53 (21) 202 (79)

35 (22) 123 (78)

18 (19) 79 (81)

0.53

316 (78) 88 (22)

170 (75) 57 (25)

146 (82) 31 (18)

0.07

150 (60) 101 (40)

89 (57) 68 (43)

61 (65) 33 (35)

0.23

213 (53) 191 (47)

123 (54) 104 (46)

90 (51) 87 (49)

0.50

41 (16) 214 (84)

27 (17) 131 (83)

14 (14) 83 (86)

0.57

352 (87) 52 (13)

203 (89) 24 (11)

149 (84) 28 (16)

0.14

197 (77) 58 (23)

127 (80) 31 (20)

70 (72) 27 (28)

0.17

384 (95) 1 (0.5) 6 (1.5) 13 (3)

219 (96) 0 (0) 1 (1) 7 (3)

165 (93) 1 (1) 5 (3) 6 (3)

0.06

97 (38) 93 (36) 60 (24) 5 (2)

51 (32) 63 (40) 40 (25) 4 (3)

46 (47) 30 (31) 20 (21) 1 (1)

0.07

217 (54) 187 (46)

123 (54) 104 (46)

94 (53) 83 (47)

0.84

78 (31) 177 (69)

54 (34) 104 (66)

24 (25) 73 (75)

0.12

70 (17) 334 (83) 91% (88e94) 96% (93e97) 91% (87e931) 17% (13e21) 5% (4e8)

36 (16) 191 (84) 90% (86e95) 95% (90e97) 90% (85e93) 21% (16e28) 8% (5e12)

34 (19) 143 (81) 91% (87e95) 97% (93e99) 92% (86e95) 11% (7e16) 3% (1e7)

0.43 0.30 0.22 0.52 0.08 0.004

67 (26) 188 (74) 62% (56e69) 79% (73e83) 79% (73e83) 21% (16e26) 15% (10e21)

40 (25) 118 (75) 56% (49e65) 78% (70e83) 75% (67e81) 20% (14e27) 15% (9e23)

27 (28) 70 (72) 72% (63e82) 81% (70e87) 85% (76e91) 22% (15e32) 15% (9e25)

0.66 0.003 0.39 0.024 0.39 0.17

2

2

Abbreviations: HPVþ: human papillomavirus-related; HPV: human papillomavirus-unrelated; Zubrod PS: Zubrod Performance Scale; OS: overall survival; LRC: locoregional control; DC: distant control; PEG: percutaneous endoscopic gastrostomy; RT: radiotherapy; IMRT: intensity-modulated radiotherapy; CI: confidence interval. Note: Significant P values are in bold.

A. Spreafico et al. / European Journal of Cancer 67 (2016) 174e182

Number of cases Mean cisplatin dose Follow-up (year), median (range) Age, median (range) Gender Female Male Zubrod PS 0 1 Smoking 10 pack-years 10 pack-years RT technique IMRT 3D conformal Disease site Oropharynx Larynx Hypopharynx Unknown primary T category T0eT2 T3eT4 N category N0eN2a N2beN3 3-Year OS (95% CI) 3-Year LRC (95% CI) 3-Year DC (95% CI) 3-Year late toxicity (95% CI) PEG dependency (at 2 years)

HPVþ, N (%)

A. Spreafico et al. / European Journal of Cancer 67 (2016) 174e182

179

Table 3 Acute treatment-related AEs leading to dose delay and/or chemotherapy omission and late toxicity profile in HPVþ and HPV patients. Cohort

Type of toxicity

HPVþ, N (%)

HPV, N (%)

Total Cisplatin 200a Cisplatin >200a Total Cisplatin 200a Cisplatin >200a (N Z 404) (N Z 227) (N Z 177) (N Z 255) (N Z 158) (N Z 97) Acute toxicity

Declined Zubrod PS Myelotoxicity Significant weight loss Mucositis/radiation dermatitis Nausea/vomiting AE-related previous delay Ototoxicity Patient declined Nephrotoxicity Allergic reaction Cardiotoxicity Grade 3e4 late Bone toxicity toxicity Ototoxicity Soft-tissue fibrosis Severe dysphagia Neurotoxicity Other

66 (16) 67 (30) 60 (15) 39 (10) 33 (8) 18 (4) 25 (6) 7 (2) 7 (2) 4 (1) 0 18 (23) 12 (15) 17 (22) 21 (27) 5 (6) 5 (6)

59 (26) 45 (20) 38 (17) 24 (11) 22 (10) 18 (8) 11 (5) 7 (3) 4 (2) 4 (2) 0 11 (22) 7 (14) 10 (20) 14 (28) 4 (8) 4 (4)

7 (4) 22 (12) 22 (12) 15 (8) 11 (6) 0 14 (8) 0 3 (2) 0 0 7 (25) 5 (18) 7 (25) 7 (25) 1 (4) 1 (4)

34 (13) 30 (12) 25 (10) 19 (7) 11 (4) 10 (4) 15 (6) 11 (4) 8 (3) 0 3 (1) 5 (8) 5 (8) 11 (19) 31 (53) 0 7 (12)

32 (20) 27 (17) 17 (11) 9 (6) 9 (6) 10 (6) 10 (6) 11 (7) 7 (4) 0 3 (2) 2 (6) 3 (9) 5 (15) 19 (57) 0 4 (12)

2 (2) 3 (3) 8 (8) 10 (10) 2 (2) 0 5 (5) 0 1 (1) 0 0 3 (12) 2 (8) 6 (23) 12 (46) 0 3 (12)

Abbreviations: HPVþ: human papillomavirus-related; HPV: human papillomavirus-unrelated; Zubrod PS: Zubrod Performance Scale; AE: adverse event. a Cisplatin dose in mg/m2.

(assuming a 5% survival difference as non-inferior) (Table 5). 4. Discussion This large pooled analysis shows that cumulative cisplatin dose reduction to 200 mg/m2 is associated with inferior OS in HPV LAHNC, attributable to reduced DC. No apparent differences in OS, LRC and DC were found with cisplatin 200 versus >200 mg/m2 in the HPVþ population. An analysis of cisplatin dose 200 versus >200 mg/m2 is clinically meaningful because high-dose cisplatin (100 mg/m2 per cycle) is the most common approach in many countries and the decision whether or not to give more than two cycles of cisplatin is a common clinical dilemma. In our study, 56% HPVþ and 62% HPV patients were unable to receive >200 mg/m2 cisplatin. The inherent limitations of a retrospective analysis are acknowledged. While the sample size for the entire cohort is large, power is limited in some subsets. For instance, only a trend towards lower OS was observed with cisplatin dose reduction in the HPVþ T4 or N3 subset, almost certainly related to insufficient sample size, as evident by the power calculation. Although we found a survival benefit with cisplatin >200 mg/m2 for HPV and a potential benefit for T4 or N3 HPVþ patients, we are unable to confirm the optimal dose beyond 200 mg/m2, and it is possible that there is no threshold of efficacy between 200 and 300 mg/m2 for these patients. Although oncologic outcomes were collected prospectively for the PMH cohort, the frequency of acute toxicities was recorded retrospectively.

Also, the decision to reduce or delay cisplatin dose was at the discretion of the treating oncologists according to institutional practices. It might introduce inherent biases in patient selection since frailer patients are more likely to require dose reduction, While this may introduce bias, standard tertiary cancer centre patterns of practice are reflected in the analyses since the majority (>90%) were non-trial patients after exclusion of non-cisplatin-based clinical trials. While both highdose and weekly cisplatin regimens were included in the analysis, we are unable to exclude possible differences in response to the two regimens due to the small numbers who received weekly cisplatin [13]. Finally, pooling data from two countries may have theoretically led to heterogeneity of some patient and treatment factors. In considering results from four clinical trials, Ang [14] had hypothesised that a cumulative cisplatin dose of approximately 200 mg/m2, independent of schedule, might be sufficient to yield a beneficial antitumour effect. However, this hypothesis was based on data in an era when the impact of HPVþ HNC was just commencing to confound clinical trial results. The emergence of HPVþ HNC was not understood sufficiently well to be considered in trials at that time. In contrast, a potential benefit of higher cumulative cisplatin dose was suggested in several studies that addressed mixed populations of HNC [2,4] but was not confirmed in another study confined to OPC [15]. None of these evaluations took into account tumour HPV status. The current study analysed HPVþ and HPV separately. We observed an OS benefit when cisplatin >200 mg/m2 for HPV LAHNC patients. High

180

A. Spreafico et al. / European Journal of Cancer 67 (2016) 174e182

Fig. 2. KaplaneMeier plots for overall survival in human papillomavirus (HPV)-unrelated (HPV) (A) and HPV-related (HPVþ) (B) LAHNSCC and HPVþ high risk defined by Ang et al. [5] (HPVþ N2b-N3  10 pack-year smoking) (C) and HPVþ high risk defined by O’Sullivan et al. [3] (T4 or N3) (D). Five-year OS resulted inferior OS for cisplatin 200 versus >200 mg/m2 in the HPV (44% versus 62%, P < 0.01) (A), but no difference in the HPVþ overall (83% versus 87%, P Z 0.30) (B) and HPVþ N2b-N3  10 pack-year smoking subset (90% versus 90%, P Z 0.76) (C), although a trend of higher OS was observed in the HPVþ T4 or N3 subset (76% versus 79%, P Z 0.15) (D). Hazard ratios by cisplatin dose (mg/m2) (adjusted for age, T, N, smoking, and disease site) are as follows:

   

HPV: cisplatin Z 200 versus <200: HR 0.8 (0.5e1.3), P Z 0.32; >200 versus <200: 0.5 (0.3e0.8), P Z 0.002. HPVþ overall: cisplatin: Z200 versus <200: HR 0.98 (0.5e2.0), P Z 0.96; >200 versus <200: 0.8 (0.4e1.6), P Z 0.53. HPVþ T4 or N3 subset: HR by cisplatin >200 versus 200: 0.5 (0.2e1.1), P Z 0.07. HPVþ N2beN3_>10 pack-year smoking subset: HR by cisplatin >200 versus 200: 0.92 (0.41e2.07), P Z 0.85.

cisplatin dose seems important for DC as evidenced in the updated MACH-NC meta-analysis [16] and confirmed in our current study. However, the effect of cisplatin on LRC was not significant. Albeit a retrospective analysis, this is the first study to show that cisplatin dose >200 mg/m2 is likely important for traditional HPV LAHNC but less relevant for most HPVþ patients.

The HR (0.5, 95% CI: 0.3e0.7) for OS by cisplatin dose for HPV was significant in MVA despite the more modest sample size. This would support the likelihood that this effect was a true finding and not a statistical artifact. Despite the large sample size, HPVþ LAHNC had a correspondingly similar HR (0.6, 95% CI: 0.4e1.1) to the HPV group, but was statistically non-significant. The relatively large confidence interval

A. Spreafico et al. / European Journal of Cancer 67 (2016) 174e182

181

Table 4 Multivariable analysis for OS in HPVþ and HPV patients. Multivariable analysis e entire cohort HPVþ

OS, HR (95% CI)

P-value HPV

Cisplatin dose (mg/m2) (>200 versus 200) 0.6 (0.4e1.1) 0.104 Cisplatin dose (mg/m2) (>200 versus  200) Radiation technique (IMRT versus 3DRT) 0.4 (0.2e0.8) 0.007 Radiation technique (IMRT versus 3DRT) Age (continuous by 1 year) 0.9 (0.95e1.02) 0.306 Age (continuous, by 1 year) T category (T4 versus T0eT3) 2.3 (1.3e4.2) 0.005 T category (T3eT4 versus T0eT2) N category (N2ceN3 versus N0eN2b) 2.5 (1.4e4.6) 0.002 N category (N2beN3 versus N0eN2a) Smoking pack-year (continuous by 10) 1.2 (1.0e1.3) 0.009 Smoking pack-year (continuous by 10) Disease site Not Included Disease site (OPC/CUP versus LHC) Multivariable analysis e IMRT subset (adjusted for age, smoking pack-years, T category, N category) Cisplatin dose (mg/m2) Cisplatin dose (mg/m2) a >200 versus <200 0.6 (0.2e1.4) >200 versus <200 Z200 versus <200 0.8 (0.4e1.6)b 0.460 Z200 versus <200

OS, HR (95% CI)

P-value

0.5 0.6 1.0 1.4 2.5 1.1 1.2

<0.001 0.009 0.980 0.098 <0.001 0.065 0.495

(0.3e0.7) (0.4e0.9) (0.98e1.02) (0.9e2.2) (1.6e4.2) (1.0e1.2) (0.8e1.7)

0.7 (0.4e1.2)a 0.4 (0.2e0.7)b

0.006

Abbreviations: HPVþ: human papillomavirus-related; HPV: human papillomavirus-unrelated; OS: overall survival; IMRT: intensity-modulated radiotherapy; HR: hazard ratio; CI: confidence interval; OPC: oropharyngeal cancer; CUP: carcinoma of unknown primary; LHC: laryngohypopharyngeal cancer; 3DRT: three-dimensional conformal radiotherapy. Note: T4, N2c-N3, 3DRT and higher smoking pack-years predicted inferior in HPVþ cohort. N2b-N3 and 3DRT predicted inferior OS in HPV cohort. Note: Significant P values are in bold. a Cisplatin dose Z 200 versus <200 mg/m2: HPVþ: P Z 0.21; HPV: P Z 0.16. b Cisplatin dose >200 versus <200 mg/m2: HPVþ: P Z 0.49; HPV: P Z 0.001.

suggests that the survival benefit of high-dose cisplatin may vary among different HPVþ risk groups. In fact, the power calculations suggest that similar to HPV patients, high-risk HPVþ patients (T4 and N3) might also require chemo-intensification >200 mg/m2. In contrast, low-risk HPVþ patients are unlikely to benefit from high-dose cisplatin. These findings support the hypothesis of cisplatin dose de-intensification in low-risk HPVþ LAHNC (T1-3N0-2b, <10 pack year smoking) [3] as a reasonable strategy to improve the therapeutic index and is currently under evaluation. One of the major clinical challenges in combining cisplatin with radiation remains the poor treatment tolerance. Better supportive care remains an active research area [17]. Alternative strategies are currently under investigation including incorporation of novel agents and tumour hypoxia modification. Consistent with other reports [18], we found that IMRT was an independent survival predictor regardless of tumour HPV status. The benefit of IMRT is likely a surrogate for improved RT quality due to improvement

in target definition with better imaging modalities (e.g. magnetic resonance imaging and positron-emission tomography), dosimetric coverage of targets, normal tissue avoidance and pectisee RT delivery using daily image guidance. RT quality has been shown to be a determinant of outcome in a protocol-required analysis of a randomised trial [19]. Evaluation of the technical factors contributing to the improved outcome with IMRT was not a defined objective in the current study and not further explored. 5. Conclusion A differential effect of cisplatin dose intensity reduction on OS for HPVþ and HPV LAHNC was demonstrated. Cisplatin >200 mg/m2 appears to be necessary for HPV LAHNC. High OS was achieved in HPVþ LAHNC overall in all cisplatin dose levels (<200, Z200, >200 mg/m2) suggesting that high-dose cisplatin is not universally required. The study data, albeit retrospective, emanate from two tertiary referral cancer centres in

Table 5 Power calculation for various patient subgroups multivariable analysis for OS adjusted for age, smoking pack-year and radiotherapy technique. Cisplatin >200 versus 200 mg/m2

N

HR (95% CI)

P-value

Total case required for significance

Requires cisplatin >200 mg/m2

HPV HPV HPV HPV HPVþ HPVþ

255 153 102 107 214 83

0.5 0.4 0.7 0.5 0.8 0.7

<0.001 <0.001 0.29 0.07 0.75 0.64

NA NA 200 (additional 98 cases) 148 (additional 41 cases) Non-inferiority 1650 (additional 1567 cases)

Yes Yes Likely Very likely No Maybe

Overall LHC OPC/CUP High risk (T4 or N3) Low risk (T0e3N0-2b) Intermediate risk (T0-3N2c)

(0.3e0.7) (0.2e0.6) (0.4e1.4) (0.2e1.1) (0.2e3.3) (0.2e3.0)

Abbreviations: OS: overall survival; HPVþ: human papillomavirus-related; HPV: human papillomavirus-unrelated; HR: hazard ratio; CI: confidence interval; OPC: oropharyngeal cancer; CUP: carcinoma of unknown primary; LHC: laryngo-hypopharyngeal cancer; NA: not applicable.

182

A. Spreafico et al. / European Journal of Cancer 67 (2016) 174e182

Europe and North America and reflect standard clinical practice. Although warranting further validation, to our knowledge, the findings are unique and have potentially pivotal implications to current practice and clinical trials. Future investigation of chemotherapy dose intensity in LAHNC should be analysed in HPVþ and HPV separately. HPVþ LAHNC requires further riskstratified strategies, including de-intensification in lowrisk cases but intensification for high-risk (T4 or N3) subsets. These observations are important since the evidence remains that HPV patients appear to be at risk of being undertreated if the cisplatin dose is reduced. Conflict of interest statement None declared. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-forprofit sectors.

Acknowledgements The authors acknowledge and are grateful for the support of the Bartley-Smith/Wharton, the Gordon Tozer, the Wharton Head and Neck Translational, Dr. Mariano Elia, Petersen-Turofsky Funds, and the ‘The Joe & Cara Finley Centre for Head & Neck Cancer Research’ at the Princess Margaret Foundation.

References [1] Green JA, Dawson AA, Fell LF, Murray S. Measurement of drug dosage intensity in MVPP therapy in Hodgkin’s disease. Br J Clin Pharmacol 1980;9:511e4. [2] Nguyen-Tan PF, Zhang Q, Ang KK, Weber RS, Rosenthal DI, Soulieres D, et al. Randomized phase III trial to test accelerated versus standard fractionation in combination with concurrent cisplatin for head and neck carcinomas in the Radiation Therapy Oncology Group 0129 trial: long-term report of efficacy and toxicity. J Clin Oncol 2014;32:3858e66. [3] O’Sullivan B, Huang SH, Siu LL, Waldron J, Zhao H, PerezOrdonez B, et al. Deintensification candidate subgroups in human papillomavirus-related oropharyngeal cancer according to minimal risk of distant metastasis. J Clin Oncol 2013;31:543e50. [4] Strojan P, Vermorken JB, Beitler JJ, Saba NF, Haigentz Jr M, Bossi P, et al. Cumulative cisplatin dose in concurrent chemoradiotherapy for head and neck cancer: a systematic review. Head Neck 2016;38(Suppl. 1):E2151e8. [5] Ang KK, Harris J, Wheeler R, Weber R, Rosenthal DI, NguyenTan PF, et al. Human papillomavirus and survival of patients with oropharyngeal cancer. N Engl J Med 2010;363:24e35.

[6] O’Sullivan B, Huang SH, Perez-Ordonez B, Massey C, Siu LL, Weinreb I, et al. Outcomes of HPV-related oropharyngeal cancer patients treated by radiotherapy alone using altered fractionation. Radiother Oncol 2012;103:49e56. [7] El-Naggar AK, Westra WH. p16 expression as a surrogate marker for HPV-related oropharyngeal carcinoma: a guide for interpretative relevance and consistency. Head Neck 2012;34: 459e61. [8] Shi W, Kato H, Perez-Ordonez B, Pintilie M, Huang SH, Hui A, et al. Comparative prognostic value of HPV16 E6 mRNA compared with in situ hybridization for human oropharyngeal squamous carcinoma. J Clin Oncol 2009;27:6213e21. [9] Chung CH, Zhang Q, Kong CS, Harris J, Fertig EJ, Harari PM, et al. p16 protein expression and human papillomavirus status as prognostic biomarkers of nonoropharyngeal head and neck squamous cell carcinoma. J Clin Oncol 2014;32: 3930e8. [10] Wong K, Huang SH, O’Sullivan B, Lockwood G, Dale D, Michaelson T, et al. Point-of-care outcome assessment in the cancer clinic: audit of data quality. Radiother Oncol 2010;95: 339e43. [11] Huang SH, Xu W, Waldron J, Siu LL, Shen X, Tong L, et al. Refining American Joint Committee on Cancer/Union for International Cancer Control TNM stage and prognostic groups for human papillomavirus-related oropharyngeal carcinomas. J Clin Oncol 2015;33:836e45. [12] O’Sullivan B, Huang SH, Su J, Garden AS, Sturgis EM, Dahlstrom K, et al. Development and validation of a staging system for HPV-related oropharyngeal cancer by the International Collaboration on Oropharyngeal cancer Network for Staging (ICON-S): a multicentre cohort study. Lancet Oncol 2016;17:440e51. [13] Homma A, Inamura N, Oridate N, Suzuki S, Hatakeyama H, Mizumachi T, et al. Concomitant weekly cisplatin and radiotherapy for head and neck cancer. Jpn J Clin Oncol 2011;41: 980e6. [14] Ang KK. Concurrent radiation chemotherapy for locally advanced head and neck carcinoma: are we addressing burning subjects? J Clin Oncol 2004;22:4657e9. [15] Barkati M, Fortin B, Soulieres D, Clavel S, Despres P, Charpentier D, et al. Concurrent chemoradiation with carboplatin-5-fluorouracil versus cisplatin in locally advanced oropharyngeal cancers: is more always better? Int J Radiat Oncol Biol Phys 2010;76:410e6. [16] Pignon JP, le Maitre A, Maillard E, Bourhis J. Meta-analysis of chemotherapy in head and neck cancer (MACH-NC): an update on 93 randomised trials and 17,346 patients. Radiother Oncol 2009;92:4e14. [17] Bossi P, Russi E, Numico G, De Sanctis V, Denaro N, Ghi MG, et al. Defining best practices on supportive care during chemoradiotherapy for head and neck cancer with a formal systemic methodology. Ann Oncol 2014;25. [18] Beadle BM, Liao KP, Elting LS, Buchholz TA, Ang KK, Garden AS, et al. Improved survival using intensity-modulated radiation therapy in head and neck cancers: a SEER-medicare analysis. Cancer 2014;120:702e10. [19] Peters LJ, O’Sullivan B, Giralt J, Fitzgerald TJ, Trotti A, Bernier J, et al. Critical impact of radiotherapy protocol compliance and quality in the treatment of advanced head and neck cancer: results from TROG 02.02. J Clin Oncol 2010;28: 2996e3001.