Does Smoking Influence Outcomes in Human Papillomavirus—Associated Oropharyngeal Cancer Treated with Transoral Robotic Surgery?

Abstract

Objective

Smoking has been identified as a risk factor for recurrence in HPV‐associated oropharyngeal squamous cell carcinoma (OPSCC) treated with definitive chemoradiation. Whether the same prognostic significance holds with surgical treatment such as transoral robotic surgery (TORS) is less well studied.

Study Design

Retrospective review.

Setting

Tertiary care academic center.

Methods

Patients with HPV‐associated OPSCC treated with TORS from January 2016 to December 2023 were included. Demographics, smoking history, disease characteristics, treatment, and oncologic outcomes were collected.

Results

156 patients were included. 87.2% were male with median age 60.0 years. Seventy‐seven were never smokers, 63 former smokers, and 16 current smokers, with a median 20 pack‐years among ever smokers. Three local, 6 regional, and ten distant metastases were identified, with 6 deaths, over a median 2.2 years of follow up. When comparing never, former, and current smokers, ever smokers demonstrated more advanced pathologic nodal staging (P = .03) and extranodal extension (P = .05). When comparing never and ever smokers, never smokers were younger (P = .04), and primary site differed with borderline significance (P = .05). Ever smokers again had more advanced nodal stage (P = .01) and extranodal extension (P = .02). There were no significant differences in adjuvant treatment, disease recurrence, or death by smoking history. There were no significant differences apart from age when analyzed using a 10 or 20‐pack‐year smoking history cutoff (P = .01, P = .02).

Conclusion

Smoking history is associated with more advanced nodal stage and extranodal extension. However, smoking does not appear to influence recurrence or death in the setting of HPV‐associated OPSCC definitively treated with TORS.

Smoking is a widely recognized risk factor for head and neck cancer, including oropharyngeal squamous cell carcinoma (OPSCC). ¹ , ² , ³ , ⁴ More recently, HPV‐associated OPSCC has risen in incidence and surpassed smoking as the most common risk factors for oropharyngeal cancer. ⁵ , ⁶ Nonetheless, smoking continues to be implicated as a critical prognostic variable even in HPV‐associated OPSCC. Previous studies have demonstrated smoking history as a predictive variable used to risk‐stratify patients with HPV‐associated OPSCC treated with definitive chemoradiation. ⁷ Additional studies have also supported the finding that smoking history is associated with worse oncologic outcomes in patients with OPSCC treated with definitive radiation with or without chemotherapy. ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³

Surgical management of HPV‐associated OPSCC is an alternative to definitive chemoradiation. This alternative treatment approach has demonstrated comparable oncologic outcomes with the potential for improved functional outcomes, sparing patients the toxicities associated with definitive chemoradiation. ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ Transoral robotic surgery (TORS) is a novel surgical technique that has been widely adopted in the treatment of HPV‐associated OPSCC, offering a less invasive approach compared to traditional surgical approaches while also achieving favorable oncologic outcomes. ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²²

While smoking is demonstrated as a risk factor for recurrence in HPV‐associated OPSCC treated with definitive chemoradiation, whether smoking carries the same prognostic significance in patients undergoing surgical treatment is less well understood. Some studies have found no difference in oncologic outcomes by smoking history, ²³ , ²⁴ , ²⁵ , ²⁶ , ²⁷ while other studies have found that smoking history has a significant effect on outcomes. ²⁸ One study found that smoking history was associated with regional recurrence, yet had no significant effect on overall survival. ²⁹ Thus previous studies have arrived at inconsistent conclusions regarding the effect of smoking history on oncologic outcomes in surgically treated HPV‐associated OPSCC.

There remains limited and inconsistent reporting on the effect of smoking history on oncologic outcomes in HPV‐associated OPSCC treated with TORS. Roden et al performed a retrospective review of 258 patients with HPV‐associated OPSCC who underwent surgical treatment at a single institution, and found that smoking history was not a significant predictor of recurrence‐free survival. ²⁴ Thus this study aims to add to the limited literature exploring the prognostic significance of smoking history in patients with HPV‐associated OPSCC treated with TORS. This has important implications for the risk stratification of HPV‐associated OPSCC, particularly as many recent and ongoing de‐escalation trials exclude patients with significant smoking history under the presumption that this is associated with increased risk of recurrence. ³⁰ , ³¹ , ³²

Methods

This study was approved by the University of Maryland IRB under protocols HP‐00110800 and HP‐00099371.

All patients with HPV‐associated OPSCC treated with TORS from January 2016 to December 2023 were retrospectively reviewed. Patients were included if they had documented smoking history as well as oncologic follow up. Patients diagnosed with non‐HPV‐associated OPSCC or without follow‐up information in the electronic health record were excluded.

Demographics, smoking history, disease characteristics, treatment, and oncologic outcomes were collected from the electronic medical record. Pathologic T and N stages were extracted from surgical pathology reports. Smoking history was categorized as never, former, or current, with less than 1‐pack‐year considered never smokers. Smoking was also considered using 10 and 20 pack‐year thresholds, as these have been used in eligibility criteria for recent de‐escalation trials. ³⁰ , ³¹ , ³²

Following TORS, all patients were discussed at a multidisciplinary tumor board to determine adjuvant treatment recommendations. Patients with T3 or T4 primaries, one positive node >3 cm or multiple positive nodes, perineural invasion, or lymphovascular invasion on final pathology are typically recommended for adjuvant radiation, while those with extranodal extension are recommended for adjuvant chemoradiation. Patients with positive margins may be recommended for re‐resection if surgically feasible, or alternatively for adjuvant chemoradiation.

Statistical analyses were performed using SPSS. Categorical variables were analyzed using chi‐square test or Fisher's exact test when there were expected counts less than 5. Continuous variables were analyzed using independent samples t test or one‐way analysis of variance test.

Results

During the study period, 156 patients underwent TORS for HPV‐associated OPSCC and had smoking history data available. All patients also underwent a concurrent neck dissection at the time of their TORS. The majority of patients were males (136 [87.2%]) who identified as White (142 [91.0%]). Median age was 60.0 years (range 40.8‐80.8). Seventy‐seven (49.4%) patients were never smokers, 63 (40.4%) former smokers, and 16 (10.3%) current smokers, with a median 20 pack‐years among ever smokers. When evaluating pack‐year thresholds, 59 (37.8%) patients had a greater than 10 pack‐year smoking history, while 35 (22.4%) patients had a greater than 20 pack‐year smoking history. Patient demographics are summarized in Table 1.

Table 1.

Patient Demographics

Sex
Male	136 (87.2%)
Female	20 (12.8%)
Median age (years)	60.0 (range 40.8‐80.8)
Race
White	142 (91.0%)
African American	12 (7.7%)
Other	2 (1.3%)
Smoking history
Never	77 (49.4%)
Former	63 (40.4%)
Current	16 (10.3%)
>10 pack‐years	59 (37.8%)
>20 pack‐years	35 (22.4%)

Most patients had primary tumors of the palatine tonsil (85 [54.5%]) and base of tongue (56 [35.9%]), followed by cancer of unknown primary (7 [4.5%]), glossotonsillar sulcus (7 [4.5%]), and pharyngeal wall (1 [0.6%]). Pathologic tumor stage was T1 in 83 (53.2%) patients, T2 in 63 (40.4%), followed by T0 in 7 (4.5%) and T3 in three (1.9%) patients. Of the three patients staged as pT3, one had been clinically staged as T2, while the other two had concordant clinical staging however were felt to be appropriate candidates for TORS in the setting of exophytic primaries which appeared confined to the palatine tonsil without extension to the base of tongue. One hundred twenty‐four (79.5%) patients had pathologic N1 disease, signifying four or fewer involved lymph nodes. Nineteen (12.2%) had N0 disease, while 13 (8.3%) N3. Perineural invasion was reported in eight (5.1%), and lymphovascular invasion in 47 (30.1%). Extranodal extension was reported in 44 (28.2%). None of the patients identified to have pathologic extranodal extension had preoperative suspicion for this finding on clinical examination or imaging, reflecting previous studies which have found imaging to have only a 62‐73% sensitivity for extranodal extension. ³³ , ³⁴ , ³⁵ Furthermore, patients with pretreatment imaging suspicious for extranodal extension or with extensive nodal burden are typically recommended for definitive chemoradiation by our multidisciplinary tumor board. While 51 (32.7%) patients underwent TORS alone, 68 (43.6%) received adjuvant radiation and 37 (23.7%) adjuvant chemoradiation. Disease characteristics and treatment received are summarized in Table 2.

Table 2.

Disease Characteristics and Treatment Received

	All (n = 156)	Never smokers (n = 77)	Former smokers (n = 63)	Current smokers (n = 16)	P‐value
Primary oropharyngeal subsite					.002
Palatine tonsil	85 (54.5%)	39 (50.6%)	37 (58.7%)	9 (56.3)
Base of tongue	56 (35.9%)	34 (44.2%)	21 (33.3%)	1 (6.3%)
Cancer of unknown primary	7 (4.5%)	3 (3.9%)	3 (4.8%)	2 (12.5%)
Glossotonsillar sulcus	7 (4.5%)	1 (1.3%)	2 (3.2%)	4 (25.0%)
Pharyngeal wall	1 (0.6%)	1 (1.3%)	0 (0%)	0 (0%)
T stage					.68
T0	7 (4.5%)	2 (2.6%)	3 (4.8%)	2 (12.5%)
T1	83 (53.2%)	43 (55.8%)	33 (52.4%)	7 (43.8%)
T2	63 (40.4%)	30 (39.0%)	26 (41.3%)	7 (43.8%)
T3	3 (1.9%)	2 (2.6%)	1 (1.6%)	0 (0%)
N stage					.03
N0	19 (12.2%)	13 (16.9%)	4 (6.3%)	2 (12.5%)
N1	124 (79.5%)	62 (80.5%)	50 (79.4%)	12 (75.0%)
N2	13 (8.3%)	2 (2.6%)	9 (14.3%)	2 (12.5%)
Perineural invasion	8 (5.1%)	6 (7.8%)	1 (1.6%)	1 (6.3%)	.32
Lymphovascular invasion	47 (30.1%)	21 (27.3%)	18 (28.6%)	8 (50.0%)	.37
Extranodal extension	44 (28.2%)	15 (19.5%)	24 (38.1%)	5 (31.3%)	.05
Adjuvant treatment					.07
None	51 (32.7%)	25 (32.5%)	18 (28.6%)	8 (50.0%)
Radiation	68 (43.6%)	39 (50.6%)	23 (36.5%)	6 (37.5%)
Chemoradiation	37 (23.7%)	13 (16.9%)	22 (34.9%)	2 (12.5%)

Median follow‐up was 2.24 years (range 0.26‐7.56). Local recurrence was reported in 3 (1.9%) patients, regional recurrence in 6 (3.8%) patients, and distant metastasis in 10 (6.4%) patients. At last follow up, 150 (96.2%) patients remained alive, while 6 (3.8%) had died. Oncologic outcomes are summarized in Table 3.

Table 3.

Oncologic Outcomes

		P‐value for different comparisons
	n (%)	Never, former, vs current smokers	Never vs ever smokers	10 pack‐year threshold	20 pack‐year threshold
Local recurrence	3 (1.9%)	0.22	0.62	1.00	0.54
Regional recurrence	6 (3.8%)	0.57	0.68	0.67	0.62
Distant metastasis	10 (6.4%)	0.34	0.33	0.51	1.00
Any recurrence	15 (9.6%)	0.36	0.28	0.58	0.75
Death	6 (3.8%)	0.23	0.21	0.67	0.34

When comparing never, former, and current smokers, patient demographics including gender, age, and race did not differ significantly across these smoking groups. Primary tumor site differed significantly (P = .002) when comparing never, former, and current smoker, with the majority of glossotonsillar sulcus tumors occurring in current smokers, however, the palatine tonsil was universally most common. Never smokers demonstrated less advanced pathologic nodal staging (P = .03) and less extranodal extension (P = .05). Disease characteristics including T stage, PNI, and LVI did not differ significantly across these three smoking groups, and there was no significant difference in adjuvant treatment received. Local recurrence (P = .22), regional recurrence (P = .57), distant metastasis (P = .34), any recurrence (P = .36), and death (P = .23) did not differ significantly across smoking groups.

When comparing never and ever smokers, never smokers were significantly younger with a mean age of 58.9 years, compared to 61.9 years amongst smokers (P = .04). Gender and race did not differ significantly. Primary tumor site differed with borderline significance (P = .05), including slightly more palatine tonsil primaries amongst ever smokers and slightly more base of tongue primaries amongst never smokers. T stage, PNI, and LVI again did not differ between never and ever smokers, and there was no significant difference in adjuvant treatment received. Never smokers also once again demonstrated less advanced pathologic nodal staging (P = .01) and less extranodal extension (P = .02). Local recurrence (P = .62), regional recurrence (P = .68), distant metastasis (P = .33), any recurrence (P = .28), and death (P = .21) did not differ significantly between never and ever smokers. Outcomes among never smokers included 2.6% local recurrence, 2.6% regional recurrence, 3.9% distant metastases, and 1.3% death rates. Outcomes among ever smokers included 1.3% local recurrence, 5.1% regional recurrence, 8.9% distant metastases, and 6.3% death rates.

When analyzed using a 10‐pack‐year or 20‐pack‐year smoking history threshold, there were no significant differences apart from age (P = .01, P = .02). The differences in N stage and ENE seen when comparing categorical groups of smokers were no longer significant when considering smoking history using these numerical thresholds. There furthermore continued to be no significant difference in adjuvant treatment received, disease recurrence, or death.

Discusssion

This retrospective review of 156 patients with early‐stage HPV‐associated OPSCC who were treated with TORS demonstrated no association between smoking history and oncologic outcomes. These results suggest that smoking history may not be significantly associated with prognostic outcomes in this patient population. This contrasts previous studies of HPV‐associated OPSCC treated with definitive chemoradiation, which identified a significant association between smoking history and poorer oncologic outcomes.

In this study, never smokers demonstrated less advanced pathologic nodal staging and less extranodal extension. Prior studies have reported a similar association between smoking status and cervical metastasis as well as extranodal extension. ²⁷ , ³⁶ Some postulated this might be due to microvascular constriction causing local ischemia and accumulation of toxic substances, in combination with impaired immune responses in smokers, which ultimately facilitate lymphatic and capsular invasion. ²⁴ , ³⁶ Additionally, smoking has been associated with distinct molecular alterations, which could promote more aggressive nodal disease. ³⁶ , ³⁷ , ³⁸ However, Roden et al did not find a significant difference in nodal staging, characterized by more than 4 lymph nodes, based on smoking status. ²⁴ Further study may be required to clarify the specific effect of smoking history on both extent as well as extranodal extension identified in cervical lymph node disease.

Importantly, there were no significant associations between smoking status and any of the oncologic outcomes studied. The prognostic significance of smoking history in patients with HPV‐associated OPSCC undergoing definitive surgical treatment has not been comprehensively and prospectively studied. Similar to the current study, Roden et al found no significant associations between smoking status and recurrence‐free survival or overall survival in a cohort of patients with HPV‐associated OPSCC undergoing upfront TORS. ²⁴ In a review of TORS outcomes for oropharyngeal carcinoma, Moore et al likewise reported no significant difference in locoregional recurrence among patients with HPV‐associated OPSCC using a 10 pack‐year smoking history cutoff. ²⁶ In another study of OPSCC treated with TORS, Stuken et al found that smoking status did not predict prognostic differences in recurrence or survival, although this study included patients with both HPV‐associated and HPV‐negative disease. ²⁷ These findings in TORS‐treated patients contrast those reported by previous studies of HPV‐associated OPSCC treated with definitive chemoradiation. Ang et al notably reported that smoking history was predictive of overall survival, ⁷ and this has been corroborated by multiple subsequent studies in the definitive chemoradiation setting. ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² This difference may suggest that patients with OPSCC and a smoking history could potentially have improved outcomes with surgical management over definitive chemoradiation, however, further studies are needed to explore this possibility.

Given that patients with significant smoking history are currently excluded from many clinical trials, including those studying de‐escalated treatment regimens for HPV + OPSCC, ³⁰ , ³¹ , ³² clarifying the true prognostic value of smoking in this population is critical. This study suggests that smoking history may not be associated with worse oncologic outcomes in patients with HPV‐associated OPSCC treated with upfront TORS, and thus these patients may not warrant exclusion from de‐escalation trials.

The majority of the study population were White males. This is consistent with previous studies, ⁷ , ²⁴ , ³⁹ , ⁴⁰ and may be due at least in part to the epidemiology of HPV‐associated OPSCC, but nonetheless may limit extrapolation of conclusions to other populations. The palatine tonsil and base of tongue are known as the most common sites of OPSCC, ⁵ and this was consistent in our patient population. Most patients in this study had T1 to T2 tumors. The predominance of early‐stage primaries of the palatine tonsil was likely due in part to the fact that included patients had been selected for surgical management via TORS. ²⁰ Relative to some previous studies of HPV‐associated OPSCC treated with TORS, ²⁴ , ²⁵ , ²⁶ , ²⁷ more patients underwent TORS alone (32.7%) and fewer patients underwent TORS with adjuvant chemoradiation (23.7%) in this cohort. This may reflect improved patient selection for TORS with time, and an increased effort to avoid adjuvant chemoradiation therapy when possible. Recurrence rates reported in this study appear comparable to those of previous studies. ²⁴ , ²⁵ , ²⁶

With regards to categorical stratification of smoking history, this study included slightly more never smokers and fewer current smokers when compared to some previous studies of patients with HPV‐associated OPSCC. ⁷ , ²⁴ , ⁴⁰ However, there are other previous studies with a nearly identical breakdown of smoking history to that of the current study. ²⁶ With regards to smoking history considered using a numerical threshold, the 37.8% rate of greater than 10 pack‐year smoking history in this study falls within the 33.4‐60% range reported by previous studies. ²⁴ , ²⁶ , ²⁷ , ⁴⁰

Limitations to this study include its retrospective nature. Only patients with reported smoking history were included, with no ability to account for those who did not have this detail documented in the electronic medical record. Future studies may benefit from a prospective template built into the electronic medical record for collection of both a numerical pack‐year smoking history, as well as current smoking status. As mentioned previously, the majority of included patients were White males, limiting extrapolation to other patient groups. Finally, future studies may also benefit from multi‐institutional collaboration, which may provide larger sample sizes with greater power to identify significant differences between patient groups with variable smoking histories.

Conclusion

In this cohort of patients with HPV‐associated OPSCC definitively treated with TORS with or without adjuvant radiation and chemotherapy, smoking history was not significantly associated with locoregional recurrence, distant metastasis, or overall survival. Smoking history was associated with more advanced nodal stage and extranodal extension.

Author Contributions

Jane Y. Tong, concept and design, acquisition, analysis, and interpretation of data, drafting of the manuscript, critical revision of the manuscript for important intellectual content, accountability for all aspects of the work; Vera Bzhilyanskaya, acquisition, analysis, and interpretation of data, drafting of the manuscript, critical revision of the manuscript for important intellectual content, accountability for all aspects of the work; Jason K. Molitoris, critical revision of the manuscript for important intellectual content, accountability for all aspects of the work; Matthew J. Ferris, critical revision of the manuscript for important intellectual content, accountability for all aspects of the work; Ranee Mehra, critical revision of the manuscript for important intellectual content, accountability for all aspects of the work; Kelly Moyer, critical revision of the manuscript for important intellectual content, accountability for all aspects of the work; Jeffrey Wolf, critical revision of the manuscript for important intellectual content, accountability for all aspects of the work; Rodney Taylor, critical revision of the manuscript for important intellectual content, accountability for all aspects of the work; Kyle M. Hatten, concept and design, critical revision of the manuscript for important intellectual content, accountability for all aspects of the work.

Disclosures

Competing interests

None.

Funding source

None.