Clinical Workup

There is no particular evidence in the medical literature guiding initial clinical evaluation, which thus relies on “common sense–based medicine.” The clinical examination is today most often performed using fiberoptic laryngoscopy or a rigid endoscope, but no study has ever prospectively compared mirror laryngoscopy (by experienced physicians) with these technologies. Evaluation under general anesthesia is performed systematically by most teams, but then again, there are no studies to “prove” that this is better than not doing it. Common sense shows that general health and comorbidities should also be thoroughly evaluated.

Laryngeal mobility is a main issue in glottic cancer. Dr Kirchner’s seminal study of whole-organ sections has shown that decreased vocal fold motion may be caused simply by a bulky tumor, but also by a tumor invading the paraglottic space. Laryngeal mobility was the only predictor of minor thyroid cartilage invasion by T1 to T3 tumors treated with conservation laryngeal surgery and for early-stage to mid-stage tumors involving the anterior commissure (level 3 evidence).

The anterior commissure (AC) must be thoroughly evaluated clinically, as the approach and outcomes differ from tumors without AC involvement (see later in this article). Subglottic extension and proximity of the tumor to the cricoid cartilage must be ascertained in view of organ-preservation surgery, in which a stable cricoid must be preserved.

Radiologic Assessment

Locoregional assessment of glottic cancer relies on computed tomography (CT), magnetic resonance imaging (MRI), and 18-fluorodeoxyglucose positron emission tomography combined with CT scan (PET/CT). CT and MRI have been shown to improve diagnostic accuracy for laryngeal carcinoma as compared with the clinical and endoscopic workup alone (level 2 evidence). Using CT, diagnostic accuracy improved from 58% to 80%, and using MRI, accuracy improved from 58% to 88%. The difference between adding CT versus MRI was not significant. Using pathology as the gold standard, reported sensitivities of CT scan for predicting cartilage invasion by laryngeal carcinoma range from 46% to 67% and can be as low as 10% for early-stage to mid-stage tumors amenable to conservation laryngeal surgery (level 3 evidence). Reported specificities range from 87% to 94% (level 3 evidence). For the diagnosis of cartilage invasion, MRI has been shown to be significantly more sensitive than CT (respective sensitivities of 89% vs 66%) but also significantly less specific than CT (respective specificities of 84% vs 94%) (level 2 evidence). Thus, there is no evidence favoring CT over MRI for the initial staging of laryngeal cancer, and each imaging modality has its limitations and pitfalls.

In the evaluation of the neck, CT, MRI, ultrasound, and PET are clearly more sensitive and specific than neck palpation alone for the diagnosis of metastatic lymphadenopathy (level 2 evidence). There does not seem to be a significant difference among these modalities in terms of sensitivity or specificity, although one study (level 2 evidence) found that MRI was more accurate for metastatic nodes smaller than 10 mm, but found no difference between MRI or CT for larger nodes. PET/CT is more accurate than PET alone for the staging of the neck (level 2 evidence). One prospective study comparing CT, MRI, ultrasound, and PET/CT using pathology as the gold standard found that PET/CT was significantly the most sensitive and specific imaging modality for detecting metastatic nodes in head and neck cancer (level 2 evidence).

Clinical Workup

There is no particular evidence in the medical literature guiding initial clinical evaluation, which thus relies on “common sense–based medicine.” The clinical examination is today most often performed using fiberoptic laryngoscopy or a rigid endoscope, but no study has ever prospectively compared mirror laryngoscopy (by experienced physicians) with these technologies. Evaluation under general anesthesia is performed systematically by most teams, but then again, there are no studies to “prove” that this is better than not doing it. Common sense shows that general health and comorbidities should also be thoroughly evaluated.

Laryngeal mobility is a main issue in glottic cancer. Dr Kirchner’s seminal study of whole-organ sections has shown that decreased vocal fold motion may be caused simply by a bulky tumor, but also by a tumor invading the paraglottic space. Laryngeal mobility was the only predictor of minor thyroid cartilage invasion by T1 to T3 tumors treated with conservation laryngeal surgery and for early-stage to mid-stage tumors involving the anterior commissure (level 3 evidence).

The anterior commissure (AC) must be thoroughly evaluated clinically, as the approach and outcomes differ from tumors without AC involvement (see later in this article). Subglottic extension and proximity of the tumor to the cricoid cartilage must be ascertained in view of organ-preservation surgery, in which a stable cricoid must be preserved.

Radiologic Assessment

Locoregional assessment of glottic cancer relies on computed tomography (CT), magnetic resonance imaging (MRI), and 18-fluorodeoxyglucose positron emission tomography combined with CT scan (PET/CT). CT and MRI have been shown to improve diagnostic accuracy for laryngeal carcinoma as compared with the clinical and endoscopic workup alone (level 2 evidence). Using CT, diagnostic accuracy improved from 58% to 80%, and using MRI, accuracy improved from 58% to 88%. The difference between adding CT versus MRI was not significant. Using pathology as the gold standard, reported sensitivities of CT scan for predicting cartilage invasion by laryngeal carcinoma range from 46% to 67% and can be as low as 10% for early-stage to mid-stage tumors amenable to conservation laryngeal surgery (level 3 evidence). Reported specificities range from 87% to 94% (level 3 evidence). For the diagnosis of cartilage invasion, MRI has been shown to be significantly more sensitive than CT (respective sensitivities of 89% vs 66%) but also significantly less specific than CT (respective specificities of 84% vs 94%) (level 2 evidence). Thus, there is no evidence favoring CT over MRI for the initial staging of laryngeal cancer, and each imaging modality has its limitations and pitfalls.

In the evaluation of the neck, CT, MRI, ultrasound, and PET are clearly more sensitive and specific than neck palpation alone for the diagnosis of metastatic lymphadenopathy (level 2 evidence). There does not seem to be a significant difference among these modalities in terms of sensitivity or specificity, although one study (level 2 evidence) found that MRI was more accurate for metastatic nodes smaller than 10 mm, but found no difference between MRI or CT for larger nodes. PET/CT is more accurate than PET alone for the staging of the neck (level 2 evidence). One prospective study comparing CT, MRI, ultrasound, and PET/CT using pathology as the gold standard found that PET/CT was significantly the most sensitive and specific imaging modality for detecting metastatic nodes in head and neck cancer (level 2 evidence).

What is the Evidence for the Optimum Therapy for In Situ Glottic Carcinoma (Tis)?

Surgery (vocal fold “stripping,” transoral laser resection, or open surgery) and radiation therapy have been widely used in the treatment of glottic Tis.

• •
  

  Initial local control ranges from 56% to 92% with surgery and from 79% to 98% with radiation therapy.

  
  
• •
  

  The final local control after salvage of recurrences ranges from 90% to 100% for both modalities.

  
  
• •
  

  Ultimate laryngeal preservation ranges from 85% to 100% for surgery and 88% to 98% for radiation therapy (level 4 evidence).

Retrospective comparative studies have shown that ultimate local control and ultimate laryngeal preservation are not significantly different between these 2 modalities (level 3 evidence). Nguyen and colleagues, however, found a significantly higher local recurrence rate after vocal fold stripping, efficiently managed with repeat surgery or radiation therapy. Le and colleagues found that involvement of the AC by the tumor was a significant factor lowering local control, using any treatment modality (level 3 evidence).

Current evidence does not show a difference in terms of ultimate oncological outcomes for Tis, but aspects other than the statistical evidence may be taken into account when treatment planning. Both transoral laser resection and radiation therapy are well tolerated, with low morbidity ; however, local possibilities and expertise play a role in treatment choice. The duration of radiation therapy and the indirect costs also intervene. Finally, radiation therapy is a “one-shot” treatment that cannot be repeated, and some have suggested that it should be used only after other modalities have failed.

What Does the Evidence Say Is the Optimal Treatment for Mid-Vocal Fold T1a Carcinoma?

As for Tis, radiation therapy and surgery, especially transoral laser resection, are widely used in the treatment of T1a glottic carcinoma. Open surgery may be an option in rare selected cases, but has been largely supplanted by transoral laser resection owing to the low morbidity.

• •
  

  Initial local control rates with both treatment modalities range from 85% to 100% (level 4 evidence). Initial local control, ultimate local control, and survival have not been found to be significantly different (level 3 evidence).

Two retrospective studies comparing contemporary cohorts (level 3 evidence) found that ultimate laryngeal preservation rates were higher for tumors initially managed surgically, as compared with initial radiation therapy:

• •
  

  96% versus 82% for Stoeckli and colleagues

  
  
• •
  

  95% versus 77% for Schrijvers and colleagues

However, a relatively recent meta-analysis of 7600 pooled patients found no significant difference in local control or larynx preservation between transoral laser surgery and radiation therapy (level 3 evidence). Thus, if local control and survival are the goal, both therapeutic options are valid, although relatively low-level evidence suggests that the ultimate laryngeal preservation rate is slightly lower for patients initially treated using radiation therapy.

Other factors determining treatment choice are cost, treatment availability, local expertise, and voice quality. Level 3 evidence suggests that transoral laser resection is less costly than radiation therapy. Transoral laser resection requires a laser and a surgical team with experience in this type of minimally invasive surgery, however, and may not be available at all sites. Radiation therapy has the “reputation” of better preserving voice quality; however, high-level evidence to prove this is lacking. Current low-level evidence is based on retrospective studies (level 4 evidence) that show conflicting results in terms of voice quality, some showing a better voice after transoral laser resection, others showing a better voice after radiation therapy. Ultimate voice quality may be determined by factors other than treatment modality, such as tumor volume or depth of tumor invasion, reflected in the different types of cordectomy in the European Laryngological Association’s classification for cordectomies. Depth of invasion may constitute a bias in some studies regarding voice, but also possibly regarding oncological outcomes. Finally, the long-term effects of treatment and the possibility of metachronous second primary head and neck cancer in these patients should be considered. In the study by Holland and colleagues, after a median follow-up of 68 months, 21% of the patients with early laryngeal cancer treated by radiation therapy developed a second primary head and neck cancer (level 4 evidence). The American Broncho-Esophagological Association recommends favoring surgery when possible for younger patients, to “save” radiation therapy as a future treatment option (level 5 evidence).

What Evidence Can Guide Treatment for Tumors T1b or T2 Involving the AC?

AC involvement by early-stage tumors has been shown by level 3 studies to be a factor for decreased local control as compared with tumors without AC invasion, whether treated surgically or with radiation therapy. Few studies, however, have directly compared these 2 treatment modalities for AC tumors.

• •
  

  In 2 studies (level 3 evidence), initial local control was better using open surgery than using radiation therapy as first-line treatment, although one of these studies found that the subgroup of “purely” AC tumors responded better to radiation therapy initially, but that final local control after salvage was worse as compared with initial surgery.

  
  
• •
  

  A third study also found that surgery (open or transoral laser resection) provided better initial local control and final laryngeal preservation than radiation therapy.

To date, there are no studies directly comparing open surgery for AC tumors with transoral resection for comparable tumors, and thus the choice of surgical approach is not evidence based, although, again, current tendencies are in favor of the transoral approach, because of evidence of lower morbidity as compared with open surgery (level 3 evidence).

In conclusion, low-level evidence suggests that one should favor surgery as the initial approach for these tumors; however, other factors may influence treatment choice. Local possibilities and expertise, as well as cost, may be involved. Exposure and tumor visualization are absolutely necessary for transoral laser resection and need to be evaluated before a treatment decision is made. Patient morphology is also a factor for radiation therapy; a low-lying larynx, near the thorax may complicate dosimetry. Finally, precise staging of the cartilage is important, but difficult, given the low sensitivity of CT for early-stage tumors involving the AC (level 3 evidence).

What Does the Evidence Say Is the Best Treatment for T2 Carcinoma?

For T2 tumors with normal vocal fold mobility treated with open conservation surgery, transoral laser resection, or radiation therapy, initial local control rates range from 84% to 95% (level 4 evidence).

• •
  

  Four retrospective comparative studies (level 3 evidence) comparing radiation therapy with open or transoral surgery found no significant difference in terms of local control or survival.

  
  
• •
  

  Regarding the surgical approach, one retrospective study found that local control was better with a supracricoid partial laryngectomy as compared with a vertical partial laryngectomy (level 3 evidence).

For T2 tumors with impaired vocal fold mobility, local control rates are lower than for T2 tumors with normal mobility, whether the treatment is radiation therapy, transoral laser resection, or open surgery, with local control rates falling as low as 50%. Tumors with impaired vocal fold mobility are at higher risk of minor cartilage invasion (28% histopathological invasion, in one retrospective study), which is often missed on pretherapeutic CT evaluation (level 3 evidence).

Even among tumors with normal vocal fold mobility, not all T2 tumors are the same. For example, Peretti and colleagues divided their group of 109 cT2s into 4 different categories according to the different tumor extensions. They found that

• •
  

  Tumors with deep extension into the paraglottic space (pT3) had a much lower rate of local control, disease-free survival, and larynx preservation (17% in each case) than more superficial T2 tumors (with respective rates of 69%–100%, level 3 evidence).

  
  
• •
  

  Subglottic extension has implications different from a supraglottic extension for local treatment but also for tumor spread to neck nodes, with a higher risk of paratracheal metastases for tumors with significant subglottic extension.

Thus, there is no high-level evidence to guide treatment choices for T2 tumors. Globally, the use of open surgery has been declining, as transoral resection has taken over as the main surgical approach for conservation laryngeal surgery, but this does not mean that open surgery is not a legitimate option. The evidence shows high rates of local control and preservation of a functional larynx with open surgery, in experienced hands for selected patients. For tumors with impaired vocal fold mobility, organ-preservation surgery is generally preferable to radiation therapy alone, but there are currently no data comparing surgery with combined modality therapy (concurrent chemoradiation) for these tumors.

What Is the Evidence Regarding the Management of the Neck for Glottic Cancer Staged T1T2cN0?

Without elective treatment of the neck, nodal recurrence rates for early-stage (T1T2) glottic carcinoma are in the range of 4%. There is no evidence that elective treatment of the neck improves regional control or disease-free survival.

A recent retrospective cancer registry study analyzed the outcomes of 73 patients with pT2cN0 glottic cancer.

• •
  

  About half of the patients had undergone elective neck dissection, with occult metastatic nodes found in 10%.

  
  
• •
  

  Multivariate analysis did not find neck dissection or adjuvant treatment to be significantly related to recurrence-free or overall survival, however (level 3 evidence).

Metastatic Delphian nodes were found in 7.5% of patients with T1b or T2 cancers with AC involvement treated with supracricoid partial laryngectomy in a recent study by Wierzbicka and colleagues. Delphian node involvement was a significant prognostic factor for locoregional failure, lower larynx preservation, and lower overall survival (level 3 evidence). This evidence, however, does not confirm the necessity for neck dissection in all patients, but encourages particular vigilance only when treating this specific subtype of cancer, and may be an argument (low-level evidence) in favor of open surgery in these cases. There are currently no guidelines or high-level evidence to guide treatment of the neck for T1T2 glottic tumors, but the low rate of occult disease and regional recurrence would favor the current practice of not treating the neck electively (level 5 evidence).

What Does the Evidence Say Is the Best Treatment for Advanced-Stage Tumors (T3–T4)?

Is There Sufficient Evidence in Favor of the Use of Exclusive Chemotherapy for Glottic Cancer?

Since the introduction of platinum-based chemotherapy, and more recently with taxane-based treatments, it has become clear that glottic cancer is chemosensitive, with one-fourth to one-third of patients being complete responders and one-half to two-thirds being partial responders. Adding chemotherapy to locoregional treatments (surgery and/or radiation therapy) has been shown to improve overall survival in head and neck cancer. Recent level 1 evidence (a meta-analysis of the effects of chemotherapy by tumor site, including 3216 patients with laryngeal cancer) has shown a significant improvement in overall survival with concomitant chemoradiation, for an absolute 5-year survival benefit of 4.5%.

In light of the advantages of chemotherapy, and the high response rates, 7 published studies have investigated using chemotherapy exclusively for treating early-stage, mid-stage, and advanced-stage glottic cancer. Five of these studies were retrospective studies of complete responders after 3 cycles of induction chemotherapy (cisplatin and 5-fluorouracil), who then were allowed to decide if they preferred locoregional treatment or to pursue chemotherapy (level 4 evidence). Four studies were from the same hospital. Four of the studies included only tumors initially considered amenable to conservation laryngeal surgery. In these studies,

• •
  

  Between 29 and 65 patients were treated with exclusive chemotherapy, for a rate of local control with chemotherapy alone ranging from 54% to 72% and an ultimate larynx preservation rate ranging from 90% to 100%

  
  
• •
  

  Toxicity was acceptable, and no chemotherapy deaths were recorded, but chemotherapy was prematurely stopped in 1% of patients because of toxicity.

To date, only 2 published studies have prospectively treated all complete responders with exclusive chemotherapy (level 3 evidence). The study by Holsinger and colleagues included 30 patients with stage II to IVa glottic (n = 14) or supraglottic (n = 16) carcinoma considered amenable to conservation laryngeal surgery:

• •
  

  Eleven patients, 4 with glottic tumors and 7 with supraglottic tumors, were complete responders after 3 to 4 cycles of chemotherapy (37%), and received 3 more cycles.

  
  
• •
  

  Ten of the 11 patients had no locoregional recurrence after a median follow-up of 5 years, for a local control rate with chemotherapy alone of 91% among the complete responders.

Divi and colleagues prospectively studied 32 patients with stage III to IVb laryngeal and hypopharyngeal tumors, and

• •
  

  Four patients, 2 with hypopharyngeal cancers and 2 with supraglottic cancers, were complete responders after 1 cycle and received additional chemotherapy.

  
  
• •
  

  All of the patients had recurrences during the additional cycles: 3 in the neck and 1 locally and regionally.

For advanced-stage tumors, exclusive chemotherapy does not provide long-term locoregional control.

Thus, low-level evidence shows that highly selected patients with early-stage glottic cancer may undergo complete remission after exclusive chemotherapy; however, the initial local control rates for these highly selected patients, initially amenable to conservation laryngeal surgery, is not better than other conservation protocols using open surgery, transoral laser resection, or radiation therapy (see earlier in this article). In addition, we currently have no means of preselecting tumors that are more biologically susceptible to respond to chemotherapy, so many patients need to be treated (between 3 and 17 ) to select the few complete responders (corresponding to 5.8%–33.0% of patients). Higher-level evidence is needed before exclusive chemotherapy can become a standard of care.

What Does the Evidence Say About Follow-up for Patients Treated for Glottic Cancer?

Follow-up aims at early detection of local recurrences, regional recurrences, distant metastases, metachronous second primaries, and complications of treatment, aiming to improve oncologic outcomes by early diagnosis of cancer-related events. Routine follow-up with regular clinical examination is performed in most centers treating laryngeal cancer, and most physicians follow patients for at least 5 years. Routine screening using panendoscopy, bronchoscopy, esophagoscopy, ultrasound, CT, and PET are performed less regularly, and there is currently no consensus regarding optimum follow-up, although several guidelines have been published by various professional societies.

• •
  

  For the diagnosis of recurrence, a prospective study by Boysen and colleagues found that 76% of recurrences (all head and neck sites combined) occurred in the first 2 years following treatment, and another 11% during the third year.

  
  
• •
  

  de Visscher and Manni found that 76% of recurrences, second primaries, or metastases (cancer-related “events”) occurred within 3 years of initial treatment (level 3 evidence).

  
  
• •
  

  Similarly, prospectively collected data analyzed by Lester and Wight found that 95% of recurrences or second primaries occurred within 2.7 years for oropharyngeal primaries, 2.3 years for hypopharyngeal primaries, and 4.7 years for laryngeal primaries (level 2 evidence).

Thus, the best evidence implies that follow-up should be the most intense for the first 3 to 5 years if one is to diagnose most of the cancer-related events in this population.

Three published guidelines recommend routine clinical examination at a rate of 19 to 25 visits for the first 5 years (with a skewed distribution toward more frequent examination during the first 2 or 3 years), based primarily on physician surveys. This does not answer the question if detection at routine follow-up improves oncologic outcomes. Conflicting level 3 evidence exists regarding the outcome if an event is detected at routine follow-up as compared with events diagnosed in symptomatic patients on self-referral:

• •
  

  For de Visscher and Manni, survival was better in the group of patients whose event was detected during a routine follow-up visit, whereas

  
  
• •
  

  For Ritoe and colleagues and Boysen and colleagues, no survival difference was found between these 2 types of patients.

In their comprehensive review of the literature, Manikantan and colleagues were unable to find convincing evidence that routine chest radiographs improved detection of second primaries or metastases, often revealed by symptoms (level 3 evidence), or that detection improved survival. They found level 4 evidence that chest CT was more sensitive than chest radiograph. They concluded that “chest CT should be done in symptomatic patients.”

For the diagnosis of neck recurrence, ultrasound has been shown to be more sensitive than neck CT, which, in turn, is more sensitive than neck palpation. There is no evidence, however, regarding any improvement in survival or any cost-effectiveness of routine follow-up neck screening using ultrasound or neck CT to detect regional recurrences, as compared with routine clinical follow-up of the neck. Level 3 evidence suggests that routine CT scanning of the neck may provide evidence in favor of recurrence earlier than clinical examination, for a proportion of patients (41% in the study by Hermans and colleagues). This does not imply that earlier detection leads to better outcomes, however.

The high sensitivity (94%), specificity (82%), and negative predictive value (95%) of PET for detecting local recurrence for head and neck cancers treated by radiation therapy or chemoradiation has been confirmed by a large meta-analysis published in 2008 (level 1 evidence). There is currently no evidence showing any improvement in patient outcomes, however, by the routine use of PET. Finally, in the systematic literature review by Manikantan and colleagues, routine follow-up bronchoscopy and esophagoscopy were found to be “not warranted,” because of the cost, morbidity, and low rate of second primary tumors diagnosed (2%–6% of patients with laryngeal cancer, level 3 evidence).

Hypothyroidism, symptomatic or subclinical, in patients treated for head and neck cancer has a reported prevalence of 5% to 56%, although the actuarial incidence at 10 years may exceed 90%. There is only one prospective study of hypothyroidism in patients treated for head and neck cancer in which the median time to hypothyroidism was 8 months, and 83% of the cases of hypothyroidism occurred within 1 year of treatment (level 3 evidence). Based on these data, current guidelines recommend thyroid function studies every 6 to 12 months following treatment of head and neck cancers.