Management of Xerostomia Related to Radiotherapy for Head and Neck Cancer

The review by Kahn and Johnstone published in this issue of ONCOLOGY is comprehensive and interesting. A few points deserve emphasis, the first of which is the issue of how we should measure and report xerostomia. Accurate and reliable measurements of xerostomia are necessary in order to properly assess its severity, time course, dose-response relationships, and the efficacy of measures to protect the glands or to stimulate salivary production following irradiation. Xerostomia encompasses the objective reduction in salivary output and changes in its composition, as well as the subjective symptoms reported by the patient. Currently available measurements of xerostomia include (1) functional imaging of gland activity, (2) measurements of the salivary output, (3) observer-assessed toxicity grading, and (4) instruments assessing patient-reported evaluation of the various xerostomia-related symptoms. Xerostomia Scoring Systems
Xerostomia is a symptom; thus, the patient's reporting of its severity is the most important factor to be assessed and ranked. Some xerostomiaspecific questionnaires have been tested for their validity and reliability. In addition, a few questions related to xerostomia have been incorporated in several comprehensive head-and neck cancer-related quality-of-life (QOL) instruments. These instruments have been tested as a whole for their validity and reliability; however, the xerostomia-related questions they contain have not been tested separately. In the clinical practice of head and neck cancer therapy, observer-defined toxicity grading is prevalent. The Radiation Therapy Oncology Group (RTOG) scoring criteria for xerostomia separated the phenomenon into acute and chronic. For acute xerostomia (occurring within 3 months of the commencement of therapy), grades are defined primarily by symptoms (degree of dry mouth, thick saliva, altered taste). The chronic xerostomia scale is ranked according to the degree of mouth dryness and the response to stimulus. It is not stated whether these are observer-rated or patient-reported, and the "response to stimulus" is not defined. Grade IV (acute salivary gland necrosis or late salivary gland fibrosis) is obviously observer-rated. No validation of this grading system has been performed. The Late Effects on Normal Tissues- Subjective, Objective Management and Analytic (LENT-SOMA) scoring system for xerostomia entails subjective grading based on evaluation of dryness and whether it is debilitating or not, objective findings of mucosal moisture, management issues such as the frequency of saliva substitutes, which are highly dependent on patients' threshold to symptoms, and salivary flow relative to pretreatment flow. Arbitrary cutoff values of salivary flow rates relative to the preradiotherapy flow rates are assigned to the various grades. The Common Toxicity Criteria (CTC) grading of xerostomia abandoned the distinction between acute and chronic grades, using both functional end points and salivary flows measures, similar to LENT-SOMA scoring. We have conducted a study comparing the RTOG scale, assessed by several observers, vs patient-reported xerostomia, and correlated each scoring system with the major salivary gland flow rates. We have found that the correlation between various observers was modest and that patient- reported xerostomia was significantly better correlated with salivary output, compared with the observerrated, RTOG scale.[1] Reducing Xerostomia
I would like to expand on issues related to efforts to reduce radiotherapy- induced xerostomia through utilizing highly conformal radiotherapy techniques such as intensity-modulated radiation therapy (IMRT). Knowledge of the maximal doses that would allow preservation of the salivary flows is an important aspect of efforts to spare the parotid glands when using IMRT for head and neck cancer. The targets-either lymph nodes in the high neck at risk of metastatic disease or adjacent primary tumors in the tonsil or nasopharynx-are very close to the deep lobes of the parotid glands. In many cases, the need arises for a compromise between lower parotid gland doses and full irradiation of the adjacent targets. If we knew how much dose we could deliver to the parotid glands without affecting their long-term function, our decisions regarding these compromises would be made with higher confidence. Data about dose-response in the parotid glands is accumulating, and some of these results are presented in the review by Kahn and Johnstone. The common finding in all these data is that a relationship seems to exist between the mean doses to the glands and their residual salivary output. The biologic implication of such a relationship is that the functional subunits in the glands are arranged in parallel, meaning that damage to some units does not necessarily cause dysfunction of the whole organ (which would be the case, for example, in the spinal cord). Rather, function will be retained until a crucial threshold number of subunits are damaged, beyond which functional deterioration is observed. Organs thought to behave in a similar way are the lungs and the liver. Beyond this finding, it is apparent from various studies that very different mean doses have been reported as thresholds beyond which functional deficit occurs. These doses range from 20 Gy to almost 40 Gy. What is the reason for this discrepancy? Several explanations are possible. An obvious explanation relates to different methodologies in assessing salivary flows. Beyond this confounding issue, there are clinical factors other than radiation dose that affect salivary output but have not been taken into account in most studies. These factors include dehydration, which is common in patients receiving head and neck radiotherapy and reduces salivary flow. Another clinical issue is the use of various medicines found to significantly affect salivary flow rates. An additional important factor is one reported recently by Coppes et al from the University Hospital in Groningen, the Netherlands.[2] They irradiated different parts of rat salivary glands using high-precision proton radiation, and found regional differences in dose-salivary production relationships. Such regional differences are likely to exist in human parotid glands, and this may be the reason for varying results found by researchers using different radiotherapy techniques that produce different dose distributions within the glands. Further research into intraparotidean regional differences in sensitivity to radiation is required for a better understanding of the limits of radiotherapy. Regardless of the dose threshold, it is now apparent that spared glands not only partly retain salivary output, but also that the output increases over time through at least 2 years after radiotherapy.[3] In contrast, generally no improvement is seen over time following standard radiotherapy in which most of the parotid glands receive full radiotherapy doses. The increase in output with spared glands translates into improvement over time in patient-reported xerostomia (an improvement that does not occur following conventionat radiotherapy).[4] Yet another issue is the importance of the submandibular glands. These glands lie in front of lymph nodes in the high neck (jugulodigastric, or subdigastric, nodes), which are usually included in the targets for irradiation. In our experience, it is not possible to spare a substantial amount of these glands while treating both sides of the neck (which is required for all advanced head and neck cancer), resulting in no output from these glands after radiation.[2] Whether the use of, for example, proton-beam technology would help is not yet known. Kahn and Johnstone have described the experience gained by Jha et al in moving one submandibular gland to the submental space, away from the radiotherapy fields. However, this technique has not yet gained broad acceptance. IMRT vs Conventional Radiotherapy
Lastly, I would like to comment on the interpretation of published results of tumor control. In order to spare the salivary glands, IMRT treats selectively predefined targets, and tissues judged not to be at risk of harboring tumor are spared. In contrast, conventional radiotherapy typically delivers radiation to extensive parts of the head and neck. Whether IMRT causes higher rates of tumor recurrence due to failure to irradiate the targets adequately is an issue requiring careful observations. Kahn and Johnstone cite the study by Dawson et al[5] and comment that "local recurrence rate was disappointingly high (21%)." However, Dawson et al reported that the 95% confidence interval for local control in their series was 68% to 90%. This means that we can be 95% sure that the true population local control rate lies between 68% and 90%. This range includes the best locoregional control rates reported by other series of IMRT or conventional radiotherapy of similar patients, and there is therefore no justification for "disappointment." Another issue relates to patient selection. IMRT is far more complex and time-consuming compared with conventional radiotherapy. It is likely that different selection factors play a role in each institution: Patients who cannot tolerate lengthy treatment, those judged to be too sick to benefit from complex therapy, those requiring urgent start of therapy, and so forth may be selected to receive simple, conventional treatment. Also, the patient population mix may be different among different institutions with regard to tumor sites (as patients with oropharyngeal cancer are expected to fare better than those with disease at other sites), tumor histology (nasopharyngeal cancer among patients of Asian origin, commonly treated on the east and west coasts of the United States, is more likely to be undifferentiated and highly responsive to therapy compared with similar cancer in Caucasian patients, who are more commonly treated in the Midwest). These factors make any attempt to compare the results of different IMRT series-or series of IMRT cases to series of conventional radiotherapy cases-futile. In general, we can state that there is no evidence that IMRT reduces the rates of locoregional tumor control compared with conventional radiotherapy, while sparing noninvolved tissues.