There are three basic strategies in the war against head and neck squamous cell cancer. Prevention is the first (so stop drinking excessively and smoking at all if you engage in these self-destructive behaviors). Killing the cancer completely the first time it is treated is the second. The third “best” strategy, salvage of initial treatment failures, is a weak fallback position that easily could be characterized as the worst. Nevertheless, in an attempt to preserve function and cosmesis, patients with head and neck cancers sometimes accept treatment plans that have an incremental chance of local failure and diminished chance of survival to avoid what might be viewed as catastrophic and potentially unacceptable consequences of radical treatment for cure. This trade-off is supported by the concept that a treatment failure can be salvaged with curative intent at a cost of still acceptable cosmetic and functional losses. All physicians who treat head and neck cancer realize salvage comes at a considerable cost in terms of increased morbidity and markedly diminished chances of disease-free survival at most primary sites.
The larynx is an interesting exception to the dim prospects of successful salvage therapy. A primary site recurrence there, discovered while the tumor is confined to the larynx, is quite likely to be cured by salvage laryngectomy. The laryngeal recurrence is still in a “box” that can be removed with an oncologically sound surgical procedure that is relatively easy to perform. Surgical salvage elsewhere is less certain and in some cases (eg, nasopharyngeal carcinoma) essentially impossible because of difficulties in clearance of tumor margins. When tumor recurs, it will be admixed with scar in otherwise deformed tissues and may be biologically more aggressive and apt to spread along nerves and vessels. Failure at the primary site then must be discovered early, before its margins become untreatably diffused into surrounding tissues, if there is to be a reasonable hope for successful salvage.
Since the early 1990s, radionuclide studies have had a recognized potential in the detection of recurrent cancer. In this issue of the AJNR, Mukherji et al (page 1215) confirm the value of another tool in the fight to improve salvage of recurrent head and neck cancer. Their study clearly supports the work of others suggesting that thallium-201 single-photon emission CT (SPECT) imaging is useful for identifying multiple head and neck lesion sites and seems to be a reasonable alternative to the less available positron-based imaging agents. This is an important contribution, for even with coincidence techniques and hardware improvements in SPECT imaging of positron emitters, studies such as 18F-flurodeoxyglucose-SPECT are more difficult to perform logistically and are more expensive than those techniques using thallium-201. This tool can and should be used selectively, as the article suggests.
Typically Mukherji et al report greater accuracy of radionuclide techniques for confirmation of recurrence as compared to CT. Much of this is owing to the lack of a post-therapy CT baseline and “conservative” CT interpretation. These points are illustrated in Figure 2 (page 1218) in which a normal asymmetry in the palatine tonsil is read as positive on the CT study. A lack of change from baseline or acceptance of this common normal variation would have improved the accuracy of CT in the study. Nevertheless, their experience and criteria likely match or exceed that generally available, so that their report is accurate within the bounds of their experimental design and the general practice of oncologic imaging in head and neck cancer patients.
Another problem with the detection of recurrent head and neck squamous cell cancer is highlighted in the article. The investigators chose to study only those patients with clinical symptoms of recurrence. Limiting an investigation to this population often means that the patient is less likely to receive salvage therapy than those whose recurrence is detected before becoming symptomatic. Pain, a common symptom of recurrence, may be caused by ulceration. Pain also can arise from disease invasion into nerves, deep musculature, and bone. In other words, symptoms often manifest late. Ulceration or growth of a mass at the primary site is often the tip of an untreatable iceberg. The authors correctly underscore that their article does not address surveillance as an attempt to detect asymptomatic recurrence. These are the people we are far more likely to identify for salvage therapy successfully.
Now that we understand the tools available for detecting asymptomatic recurrence to include CT, MR imaging, and at least two radionuclide techniques, we must begin to apply these as part of our “third-best strategy” in the war against head and neck cancer. Patients at moderate or high risk for recurrence can be identified based on pretreatment imaging and clinical criteria. These patients should have baseline post-treatment studies about 3 months after completion of therapy. The choice of study at this time is not clear, but it seems that either CT or radionuclide studies should suffice. There is no reliable reported experience with MR imaging to date regarding this issue, but it should work as well. It is very likely that a CT baseline study, showing normal post-treatment changes, will correlate with a 90% to 95% chance of local control. A second follow-up study 6 months after completion of initial therapy, showing stable post-treatment changes or no significant focal accumulation of tracer on CT, will raise the likelihood of local control to nearly 100%. The corollary of this suggested scenario is that progressive imaging changes or focal tracer accumulation will indicate recurrence in about 75% or more of the cases in which those findings occur. With such high positive and negative predictive values, biopsy, with its attendant risks, should be performed only when confronted with a very high likelihood of recurrent tumor. This group will have been triaged on the basis of a pretreatment risk profile and objective post-treatment surveillance studies.
This is not a plea for imaging all patients treated for head and neck cancer. Such a suggestion would be economically irresponsible. It is a plea for the logical and judicious application of powerful imaging tools to help improve the salvage rate and reduce the morbidity of treatment for recurrent head and neck squamous cell cancer. This will become more important as targeted nonsurgical salvage therapies become more widely available in the near future.