A report from Sher et al. published in 2011 details the results of 42 patients treated with IMRT, either in the postoperative or definitive setting. Patients were treated between 2004 and 2009 and consisted of 12 % stage I, 24 % stage II, 33 % stage III, and 31 % stage IV. Thirty of 42 patients received surgery prior to radiotherapy. The remaining 12 received concurrent chemoradiotherapy (CRT), with or without induction chemotherapy. In this population, results from definitive RT were significantly inferior. Two-year OS was 85 % and 63 %, and LRC was 91 % and 64 %, for surgical vs. nonsurgical management, respectively [23]. A larger series assessing the impact of location of primary site within the oral cavity utilized definitive CRT alone. One hundred fifteen patients, staged IIIA, IVA, and IVB (6 %, 47 %, 47 %), received a median RT dose of 72 Gy. Response rate was 96.5 % (76.5 partial, 20 % complete). Overall, 3-year OS and PFS rates were 22 and 25 %. Interestingly, the 3-year PFS rates varied notably between subsites. Tumors of the buccal mucosa, FOM, and gingiva had a 3-year PFS of 51 %. Tumors of the retromolar trigone/hard palate and tongue/lip had significantly worse PFS, at 18 % and 6 %, respectively [24]. Cohen et al. retrospectively reviewed T4 OC tumors treated with definitive CRT on prospective protocols. Thirty-nine patients were reviewed and treated with a median 74 Gy. Results at 5 years of OS, PFS, and LC were 56 %, 51 %, and 75 %, respectively. The authors concluded that these were comparable results to historical controls and that primary CRT should be considered in this population [25]. Gore et al. retrospectively reported on 104 patients with advanced OCSCC treated with curative surgery or chemoradiation and found significantly improved survival in those treated with surgery [26].

Management of the elective nodal volumes in HNSCC is an important issue. Generally, structures of the oral cavity drain to bilateral lymphatics, with preferential involvement of certain levels [27]. There are exceptions, however, leading to instances in which unilateral neck radiation may be considered if the suspected contralateral failure rate is sufficiently low. This paradigm is well studied in the setting of the palatine tonsil [28, 29]. A small study assessing well-lateralized, early-stage OC tumors and tonsillar tumors following surgery revealed a 5 yr LRC of 100 % [30]. The best example of a well-lateralized structure within the oral cavity is the buccal mucosa. In a retrospective review of 145 patients receiving surgery and adjuvant RT, 83 % received unilateral neck RT. For all patients, the 5-year disease-specific survival (DSS) for stage I–IV was 87 %, 83 %, 61 %, and 60 %, respectively. There was no difference in LRC between patients who received unilateral vs. bilateral treatment (p = 0.95). The rate of failure within the contralateral neck was only 2.1 % [31]. Another retrospective review from Vergeer et al. assessed well-lateralized tumors of the oral cavity (oral tongue, FOM, alveolar process, buccal mucosa) and oropharynx. Sixty-one percent of those included had tumors of the gingiva or buccal mucosa. Eighty-three percent of patients had ipsilateral-only neck dissection. Contralateral neck failure occurred in 6 % of patients, and the only significant prognostic factor affecting contralateral neck failure was number of positive lymph nodes. Five-year rates of contralateral neck control were 99 %, 88 %, and 73 % in N0, N1 or N2a, and N2b cases, respectively. There was no significant difference in tumor subsite related to rate of contralateral neck failure [32].

Historically, radiotherapy has been fractionated in doses of 1.8–2.0 Gy/day, for a period of several weeks. This schema allowed dose escalation to the point of adequate tumor control, while taking advantage of normal tissue’s ability to self-repair in between fractions. This pattern of treatment is generally referred to as conventional fractionation. Various alternate regimens have been attempted.

The seminal American trial for altered fractionation in squamous cell carcinoma of the head and neck (HNSCC) was Radiation Therapy Oncology Group (RTOG) 90–03. This trial enrolled about 1100 locally advanced patients (stage II–IV) with tumors of the oral cavity, oropharynx, hypopharynx, and supraglottic larynx. OC patients represented just over 10 % of the enrollees. Patients were randomized to one of four arms: (1) standard fractionation at 2 Gy/fraction/day, 5 days/week, to 70 Gy over 7 weeks; (2) hyperfractionation at 1.2 Gy, twice daily, 5 days/week to 81.6 Gy over 7 weeks; (3) accelerated fractionation with split at 1.6 Gy/fraction, twice daily, 5 days/week to 67.2 Gy over 6 weeks, including a 2-week break after 38.4 Gy; or (4) accelerated fractionation with concomitant boost at 1.8 Gy/fraction/day, 5 days/week and 1.5 Gy/fraction/day to a boost field as a second daily treatment during the last 12 days to a total of 72 Gy over 6 weeks. The trial was initially reported in 2000 [33] and revealed improved locoregional control compared with standard fractionation. While there was a trend toward improved PFS in these groups, neither it nor OS was significantly improved. All three altered fractionation groups had significantly worse acute toxicity. Final results from the trial were published in 2014 and revealed that only hyperfractionation improved LRC (HR 0.79, p = 0.05) and OS (HR 0.81, p = 0.05) without increasing delayed toxicity [34]. Another large, phase III trial was conducted in multiple continents to compare accelerated fractionation to standard fractionation. Nine hundred patients with HNSCC of the larynx, pharynx, or oral cavity were enrolled (24 % oral cavity). Standard fractionation consisted of 66–70 Gy in 33–35 fractions, while accelerated fractionation was the same total dose, while receiving 6 fractions/week. Five-year LRC was 42 % vs. 30 %, favoring the accelerated group (p = 0.004) at the cost of significantly worse mucositis and skin reaction [35].

Regarding definitive management, concurrent chemotherapy and radiation have taken the forefront in advanced-stage disease of non-oral cavity subsites. Data suggests that modifying the fractionation style of radiation cannot compensate for the benefit that concurrent chemotherapy provides when it is required [36]. Currently, with chemotherapy, the usual radiotherapy fractionation is 5–6 fractions per week to a total dose between 66 and 72 Gy in 2–2.2 Gy daily fractions.

As surgery tends be the primary option for OCSCC, not definitive RT, altered fractionation is not as frequently discussed. There have been attempts to shorten the treatment time in the postoperative setting, however. The long-term results of small pilot study from Trotti et al. suggested that locoregional control could be improved with accelerated RT, consisting of 1.8 Gy per day in just over 5 weeks, using a concomitant boost approach. The total dose was 63 Gy, and patients received a second daily fraction once a week for the first 4 weeks, then on the final four treatment days. As expected, this approach did result in increased acute mucosal reactions, but did not affect rates of long-term toxicity [37]. Another small prospective study of accelerated RT following curative resection for HNSCC has been performed [38]. While thought provoking, these deviations from conventional fractionation have not garnered great support, and standard fractionation remains standard of care in the adjuvant setting.

RTOG 73–03 evaluated preoperative versus postoperative radiotherapy. It included 277 patients, and with 10-year median follow-up, LRC was significantly better in the postoperative population; however, there was no difference in survival [39]. This study established the postoperative treatment paradigm for HNSCC. A more modern series from Germany evaluated 151 patients specifically with OCSCC and N2 disease, who underwent either pre- or postoperative chemo-RT. They found an increased 5-year survival in those who received the therapy prior to surgery (46 % vs. 27 %, p = 0.035). When broken down by T-stage, the benefit was found to be confined to patients with T4 disease [40]. This series suggests a potential subgroup of patients who may benefit from neoadjuvant CRT, perhaps due to its ability to assist in curative resection with negative margins. This hypothesis is further supported by a retrospective study assessing pathologic response rate and effect on prognosis after neoadjuvant CRT. In this series of 154 patients, a clinical response rate of nearly 93 % and a pathologic complete response rate of 60 % were observed after a preoperative RT dose of only 40 Gy, together with platinum-based chemotherapy. After a complete clinical response, residual disease was found on dissection in only 8 % of cases, and only in levels IB-IIA. The authors suggest that neoadjuvant CRT might allow for a more tailored surgical intervention, perhaps avoiding multilevel selective neck dissection [41]. In a large literature-based meta-analysis of 2015 patients (predominantly stage III–IV), complete histopathological response was found in 48 %. After a number of analyses, the mean survival rates at 3 and 5 years were 73 and 57 % [42]. These are numbers that compare favorably to cohorts treated with postoperative therapy. It remains an attractive treatment approach; however, prospective randomized trials are required to firmly establish efficacy.

Chemotherapy prior to definitive surgery for HNSCC (a.k.a. induction) has been studied for many years. Induction has also been studied extensively prior to definitive CRT in other cancers of the head and neck. This approach has many supporters and is backed by a wealth of hypothetical advantages. These facts aside, there is a lack of high-level evidence supporting the use of this modality. Large phase III trials have concluded the optimal regimen to be TPF (docetaxel, cisplatinum, 5-fluorouracil) [43, 44]. Recently, a phase III trial comparing induction chemotherapy followed by surgery to up-front surgery was reported in the OCSCC population. This trial randomized patients to two cycles of TPF followed by curative resection and postoperative RT vs. surgery and postoperative RT alone. Two hundred fifty-six patients were enrolled, and 222 completed the full course of treatment. Clinical response was 81 %. After a median of 30 months, there was no difference in OS or DFS. Again, clinical response was found to be predictive [45]. Induction chemotherapy prior to definitive surgery for OCSCC is not a recommendation from the National Comprehensive Cancer Network (NCCN) [46].

Locoregional recurrence represents a significant proportion of failures, and these are often in patients who have had prior radiation. Lee et al. have published their data on the subject, looking at 105 patients with recurrent HNSCC. Median RT dose was 59.4 Gy, and 2-year locoregional PFS and OS rates were 42 % and 37 %, respectively. The use of IMRT was found to be significant for decreasing LRF. Severe late complications were observed in 12 patients (11 %), which developed after a median of 6 months [47]. The RTOG conducted a prospective trial to assess the toxicity of re-irradiation with chemotherapy. There was a 7.6 % rate of grade 5 toxicity and an 18 % rate of grade 4 toxicity. Improved survival was seen in those who had an inter-treatment interval of greater than 1 year [48]. Comorbidity is also known to be a prognostic factor, as demonstrated by Tanvetyanon et al. who produced a nomogram, to predict probability of 24-month survival [49]. Despite the mentioned data, re-irradiation remains a high-risk venture, best left for clinical trials and experienced hands in high-volume centers.

Several studies have sought to define the tumor factors that portend high risk of recurrence and merit adjuvant radiation or chemoradiation therapy. In a prospective dose-finding trial from Peters et al., poor prognostic factors were identified in an attempt to tailor adjuvant RT dosing. Three hundred two patients with SCC of the oral cavity, oropharynx, hypopharynx, or larynx were enrolled, and 92 % had stage III–IV disease. Patients were stratified by risk group based on certain factors and randomized to one of three doses: 52.2 Gy, 63 Gy, or 68.4 Gy. Low risk could not receive the high dose and vice versa. Presumed risk factors for locoregional recurrence were drawn from prior reports and included site of disease, surgical margins, perineural invasion (PNI), number and location of involved lymph nodes, and presence of extracapsular extension (ECE). Patients receiving less than 54 Gy had higher primary failure rates. Only patients with ECE benefitted from a dose greater than 57.6 Gy, and ECE was a significant predictor of LRR. Other factors were found to confer increased risk when grouped in two or more. These included oral cavity primary, close/positive mucosal margins, PNI, 2 or more involved nodes, largest node greater than 3 cm, treatment delay greater than 3 weeks after surgery, and Eastern Cooperative Oncology Group (ECOG) performance status ≥2 [50]. A subsequent multi-institutional prospective trial that accrued between 1991 and 1995 enrolled 213 post-operative HNSCC patients. This trial aimed to validate the previously identified risk factors, while comparing conventional and accelerated RT in the adjuvant setting. Low-risk, intermediate-risk, and high-risk patients received no adjuvant treatment, 57.6 Gy and 63 Gy in either 5 or 7 weeks, respectively. Using the previously identified risk factors, patients were considered low risk if they had no adverse features, or intermediate risk if they had only one feature (excluding ECE). Anyone with ECE, or 2 or more other factors was considered high risk. Low- and intermediate-risk patients had a significantly improved LRC compared to high-risk patients, thereby validating the risk stratification. Notably, a decreased LRC rate was seen in high-risk patients experiencing a delay between surgery and the initiation of RT in the 7-week arm. Both LRC and OS were significantly worse in those who had RT initiated >6 weeks after the surgical date. Total treatment time of less than 11 weeks led to significantly improved outcomes [51].

The potential benefit of adding chemotherapy to further improve outcomes has also been studied. Two large phase III trials on this topic were accrued and published contemporaneously. EORTC (European Organization of Research and Treatment of Cancer) 22931 was a randomized trial that accrued 334 patients with SCC of the oral cavity, oropharynx, hypopharynx, or larynx. Eligible patients included those with T3–4 disease with negative margins, or early-stage disease with certain features (ECE, positive margins [defined as within 5 mm], PNI, lymphovascular invasion (LVI)). Oral cavity patients with nodal involvement of levels IV or V were included as well. Patients were randomized to RT with or without concomitant chemotherapy (cisplatin 100 mg/m² every 21 days). At 5 years, LRC, PFS, and OS were all significantly improved with the addition of chemotherapy [52]. The second trial that ran in parallel was an American one, run by the RTOG. RTOG 95–01 randomized 416 patients with SCC of the oral cavity, oropharynx, larynx, or hypopharynx following gross total resection to RT alone, or RT with chemotherapy. Eligibility for this trial was slightly different and included patients with two or more lymph nodes, ECE, or positive surgical margins [defined at the inked margin]. The primary endpoint was LRC. At 2 years, LRC was significantly improved with chemotherapy (82 % vs. 72 %). DFS was also significantly improved; however, the improvement in OS did not reach statistical significance [53]. Given the substantial similarity between the two trials, a combined analysis was undertaken. This was published in 2005 and revealed that ECE and positive surgical margins were the only risk factors for which combined chemotherapy and RT provided benefit [54]. For this reason, these factors represent Category 1 indications from the NCCN (National Comprehensive Cancer Network) for the use of combined chemoradiotherapy [46]. The above data are summarized in Table 4.2.