Use of imaging tests after primary treatment of thyroid cancer in the United States: population based retrospective cohort study evaluating death and recurrence
BMJ 2016; 354 doi: https://doi.org/10.1136/bmj.i3839 (Published 20 July 2016) Cite this as: BMJ 2016;354:i3839All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
Thyroid cancer incidence is rising, with concern for overdiagnosis of low-risk disease.1 In the setting of this rise in low-risk disease over time, we found an increase in imaging after initial treatment.2 3 We then looked at the outcome variables recurrence and disease-specific survival to assess benefit versus risks.2
We used SEER-Medicare data, which primarily includes older patients but also includes younger patients on Medicare for other reasons. SEER-Medicare data are commonly used to assess management in other cancers and therefore, its strengths and limitations were not only addressed in our paper but are also well known. Although imaging was associated with identifying more recurrence, in most cases there was no improvement in disease-specific survival. Our findings emphasize an important topic: tailoring imaging. Tailoring imaging to disease severity and limiting unnecessary use should be a goal for all physicians. We want the “right” patients to get the “right” imaging. At no time did we say that “no one” with thyroid cancer needs imaging.
This study is the foundation for future work and more details are needed to ascertain which patients do and do not need specific imaging. However, this study is an important step in the right direction.
1. Vaccarella S, Franceschi S, Bray F, et al. Worldwide Thyroid-Cancer Epidemic? The Increasing Impact of Overdiagnosis. The New England journal of medicine 2016;375(7):614-7.
2. Banerjee M, Wiebel JL, Guo C, et al. Use of imaging tests after primary treatment of thyroid cancer in the United States: population based retrospective cohort study evaluating death and recurrence. BMJ 2016;354:i3839.
3. Wiebel JL, Banerjee M, Muenz DG, et al. Trends in imaging after diagnosis of thyroid cancer. Cancer 2015;121(9):1387-94.
Competing interests: No competing interests
Dear Dr. Saripanidis:
Thank you for your interest in our comment. I am not sure that I completely understand your note, and I am particularly at a loss as to why you addressed it to us instead of the original authors of the paper under discussion. (1) You expressed an opinion that “the biggest elephant in the room is overusing and billing for neck ultrasound (NUS) in order to screen healthy patients”. As you recall, the paper under consideration did not directly address screening healthy individuals. Therefore, this proverbial elephant you identified is a mirage in this proverbial room. The actual subject you are addressing is called overdiagnosis by screening healthy individuals and/or patients that was investigated in many excellent papers directly. (2-4) The work by Vaccarella at al. may be of particular interest to you as it specifically addresses “ultrasonography in gynecologic and obstetric clinics, which favored opportunistic examinations of the thyroid gland in women of reproductive age.”(4 ) This practice in part contributed to overdiagnosis of thyroid tumors according to those authors. I wonder if that is the setting you witnessed and developed such passion for the topic.
I do agree that NUS was not “helpful in reducing mortality [in] patients already afflicted by thyroid cancers”, which one has to clarify by adding the specific circumstance of surveillance monitoring following the initial treatment. I made this point clear in my original rapid response under the paragraph that addressed the proverbial elephant #2. Another point you were making is that NUS “did not prove cost effective”. This question was not addressed in the article under consideration. (1) One has to be diligent not to extend or conflate the study findings. I find that you use unjustified liberty by addressing findings that were not even included in the study design and applying results to patients that were in fact not studied at all. The results of this retrospective cohort study must be interpreted specifically for the cohort under investigation with all the circumstantial specifics included, and applied specifically to the clinically equivalent groups of subjects. Use of NUS “to screen healthy” subjects is not even remotely equivalent to the cohort studied in the paper under the consideration and should not be extrapolated using loosely based logic.
Your suggestion that “NHS managers worldwide should use this research study to curb unnecessary healthcare expenses” is another example of ignoring specifics. Which expenses were unnecessary? If you are talking about NUS, this would be a reasonable statement based on unchallenged results of this study. However, this same paper states that PET/CT studies were not associated with survival advantage. If you read my commentary in full, I made a compelling case that invalidated this conclusion given the flaws in the study design that I pointed out. Hence, curbing PET/CT on the basis of this study is unjustified. Of course, we do not know what you think about this specific issue. In fact, if “NHS managers” take this study and curb expenses for PET/CT it may end up with jeopardizing services to seriously ill patients that desperately need them. Furthermore, I also made a compelling case that the whole body radioactive iodine surveillance scan that was associated with survival benefit may not be offered to patients who could have benefited from this life saving surveillance. I do not think you would want to have “NHS managers” curb this expense either, do you? In its present form, this study has a potential to do more harm than good. Your proclamation that this is an “excellent study” is like calling Richard Nixon an “excellent” president of the United States. Yes, he did open China for trade with the West, which was one of his few excellent accomplishments. But, as you may know, he is mostly remembered for his malfeasance in office, called the Watergate scandal, which resulted in his most shameful resignation – hardly what one would call an overall excellent career. I would encourage you to look at the specifics and you may want to reconsider the overall assessment of this paper.
References:
1. Banerjee M, Wiebel JL, Guo C, et al. Use of imaging tests after primary treatment of thyroid cancer in the United States: population based retrospective cohort study evaluating death and recurrence. BMJ 2016;354:i3839.
2. Davies L, Welch HG. Current thyroid cancer trends in the United States. JAMA Otolaryngol Head Neck Surg 2014;140(4):317-22.
3. O'Grady TJ, Gates MA, Boscoe FP. Thyroid cancer incidence attributable to overdiagnosis in the United States 1981-2011. Int J Cancer 2015;137(11):2664-73.
4. Vaccarella S, Franceschi S, Bray F, et al. Worldwide Thyroid-Cancer Epidemic? The Increasing Impact of Overdiagnosis. N Engl J Med 2016;375(7):614-7.
Competing interests: No competing interests
Dear colleague,
The biggest elephant in the room is that overusing and billing for neck ultrasound (NUS) in order to screen healthy patients, hoping to detect thyroid cancers, is like throwing $billions into trash cans, since this imaging technique did not prove cost effective and helpful in reducing mortality to patients already afflicted by thyroid cancers.
NHS managers worldwide should use this research study to curb unnecessary healthcare expenses, false positives, and all consequent morbidity, like anxiety, depression, suicides, attempted suicides, risky behaviours, injuries, biopsies, thyroidectomies, irradiation, chemotherapies, lymphadenectomies.
Dear BMJ Editors,
Please include this excellent study at your "Too much Medicine" BMJ campaign/section interactive timeline.
Competing interests: No competing interests
Dear Editor:
We read report by Banerjee et al. with concern that the authors have overlooked the three proverbial “elephants in the room”.(1) This research evaluated utilization of neck ultrasound (NUS), positron emission tomography/computed tomography (PET/CT), and radioactive iodine whole body scans (RAI WBS) in surveillance of patients with differentiated thyroid cancer (DTC). The data from the “Surveillance, Epidemiology, and End Results” (SEER) database was linked with the procedure billing codes in the Medicare (insurance for patients who are 65 years of age or older) database. Paradoxically, the lower end of age range in Table 1 is 21 (the first two columns) and 28 (the third column) years.(1) How did the young patients make it into the study? How many of them did? The authors should have addressed this in the methods.
Elephant #1: The authors concede regarding all three tests that one of the study limitations is lack of information on “patient specific data such as iodine avidity, patient preference, and indications for imaging or procedures”. Because we are professionally involved in assessing indications for all the three tests at our hospital, as well as interpreting them, it struck us that there are significant differences between the tests with respect to this issue. For instance, RAI WBS has only one indication – surveillance of DTC. Therefore, the above limitation does not apply to RAI WBS. The NUS is almost exclusively done for DTC surveillance in patients after thyroidectomy. Although there may be very rare exceptions (such as parathyroid adenoma search), they likely represent a negligible percentage. On the contrary, PET/CT is indicated in many different types of cancer and patients with pulmonary nodules (all use the same procedure codes). The test is also indicated in patients with seizures and dementia, which involves imaging only the head and coded with CPT 78810 and 78814. The authors list these codes in their methods as acceptable, but they are not applicable to DTC surveillance. Therefore, if an elderly patient in this study had clinical signs of dementia for which PET/CT was done, the authors would have counted it erroneously as a test for DTC surveillance. Furthermore, given the older age of this DTC cohort, it is likely that the study included PET/CT exams which were done for non-DTC indications. Over the period of 1999-2006 it has been shown in the Medicare population that PET/CT utilization for only the 6 most common cancers (by far not all that PET/CT is indicated for) has increased at a mean annual rate of 35.9% to 53.6%.(2) Unless the authors excluded patients with other known cancers from their database without telling us about it in their methods, it is very likely that significant number of the PET/CT examinations they counted were for non-DTC cancer indications. Therefore, PET/CT data was likely significantly compromised by the uncertainty of the primary indication. Hence, we suggest that this represents the main elephant in the room. It would be most reasonable to conclude that the rate of PET/CT utilization for DTC surveillance is not attainable from this database. If the authors disagree, they need to show that inclusion of PET/CT for unrelated indications did not significantly inflate the PET/CT counts. The poorly supported claim that PET/CT is not associated with survival advantage in DTC patients could influence payers to deny coverage, causing heightened anxiety and possibly harm in properly selected DTC patients who require it for DTC management.
Elephant #2: The authors display all three imaging tests as a combined data plotted against the years (1998-2011) in Figure 1 of the paper, and it shows a steeply rising utilization. This group previously published a report where each test was plotted separately.(3) The test done most frequently by a significant margin was NUS, and it also showed a steep growth rate.(3) In this paper, NUS was performed in 56.7% of all patients (this is more than RAI WBS and inflated PET/CT combined), and 75.6% of those patients had a subsequent procedure or treatment; however, no positive survival benefit was found with NUS according to the data.(1) Yet the authors do not advocate specifically curbing NUS, but rather explain why the test did not show survival benefit. We suggest that elephant #2, the overused NUS that has shown no survival benefit, demands specific acknowledgement by the authors in reference to curbing imaging. In the conclusion, the authors offer the nebulous truism that “findings emphasize the importance of curbing unnecessary imaging.” This general statement allows proverbial elephant #2, the overuse of NUS, to stampede in the room. Pointing out the need to curb NUS may not be popular with the USA endocrinologists who are known to perform and bill for NUS on their patients.(5)
Elephant #3: There were 23.9% of all patients (Table 1) who had RAI WBS scans. The authors find that RAI WBS was associated with disease specific survival benefit “in presumed iodine avid disease”.(1) It is estimated that about 50-80% of all patients with differentiated thyroid cancer (DTC) have iodine avid disease, which is more than twice the percent who got RAI WBS.(4) Therefore, there is likely a significant underutilization of RAI WBS in surveillance of DTC, but the comment on this possible underutilization is conspicuously absent. Ignoring possible underutilization of a survival-associated RAI WBS is the third elephant in the room that may be standing in the way of improving outcomes in DTC patients.
References:
1. Banerjee M, Wiebel JL, Guo C, et al. Use of imaging tests after primary treatment of thyroid cancer in the United States: population based retrospective cohort study evaluating death and recurrence. BMJ 2016;354:i3839.
2. Dinan MA, Curtis LH, Hammill BG, et al. Changes in the use and costs of diagnostic imaging among Medicare beneficiaries with cancer, 1999-2006. JAMA 2010;303(16):1625-31.
3. Wiebel JL, Banerjee M, Muenz DG, et al. Trends in imaging after diagnosis of thyroid cancer. Cancer 2015;121(9):1387-94.
4. Min JJ, Chung JK, Lee Y, et al. Relationship between expression of the sodium/iodide symporter and (131)I uptake in recurrent lesions of differentiated thyroid carcinoma. Eur J Nucl Med 2001;28(5):639-45.
5. American Association of Clinical Endocrinologists. Endocrine Certification in Neck Ultrasound (ECNU) program: AACE THE AMERICAN ASSOCIATION OF CLINICAL ENDOCRINOLOGISTS; 2016 [Available from: https://www.aace.com/ecnu].
Corresponding Author: Mark Tulchinsky, MD
mark.tulchinsky@gmail.com
Competing interests: No competing interests
Dear Editor:
I read with interest the paper by Banerjee & Haymart et al. which reports on the marked rise of imaging studies for thyroid cancer resulting in increased treatment for recurrence without clear improvement in disease specific survival (with the exception of radioiodine scans in “presumed” iodine-avid disease). The authors contend that this paper must be taken very seriously "as the foundation for understanding when to image and when to treat recurrence".
There are several important points that require careful consideration:
1. Using a composite of imaging tests (bundling together 131-I scans, neck US and PET/CT) to assess the trend of image utilization does not give an accurate image of thyroid cancer management. According to the data in Table 1: for the entire group of 28,220 patients, only 23.9% underwent a follow-up radioiodine (RAI) scan, and 14.9% underwent a PET scan. The imaging test that was used most frequently is Neck Ultrasound for 56.7% of patients. The authors need to also report a Mean (& range) for the number of tests done for each Imaging Modality/per patient. With this additional data, one can detect if the increased trend over time in the composite measure of imaging tests is driven by Ultrasound studies, or other imaging.
2. The authors report an association between the Follow-up RAI scans and improved disease-specific survival. In reality, the reported association with an improved disease-specific survival is due to the RAI treatments, not the follow-up RAI scans. Indeed, it is more likely that patients who received one RAI treatment also had a follow-up scan, however from Table 1 it appears that only 23.9% of patients had follow-up Radioiodine scans, while 76.1% did not.
3. It is concerning that the authors do not report how many patients underwent initial RAI treatment in the entire cohort of 28,220 patients. However, the authors determined an adjusted Odds Ratio of 17.8 for RAI Re-treatment for patients who underwent a follow-up RAI scan, without reporting how many patients underwent a second RAI treatment, how many patients underwent a third RAI treatment, and > 3 RAI treatments. Using the billing codes for therapy the authors could determine this data. Unfortunately, the authors do not even report how many patients received one RAI treatment (the initial RAI treatment). Moreover, it is unusual that the authors chose to report the indirect relationship between follow-up RAI Scan and Survival, instead of presenting the data regarding initial RAI therapy and Survival.
4. In Table 1 the authors report that 1155 (4.1%) patients died from thyroid cancer (disease-specific death): only 31% of patients who died had follow-up RAI scan, while 69% of patients did not. It is unusual that the majority of patients dying of thyroid cancer did not have a follow-up RAI scan, although RAI treatment is the mainstay of treatment and it is predicated by the findings of a RAI scan. Furthermore, regarding PET imaging, only 28% of patients who died had a PET scan, while 72% of patients who died did not. The major question is: if the majority of patients (approximately 70%) who died were not followed-up with a RAI scan, nor a PET scan, - how were they managed? This raises the question if we can really trust this data obtained from billing codes.
5. According to the data in Table 1: for the entire group of 28,220 patients, only 14.9% patients underwent a PET scan. The conclusion that PET imaging needs to be curbed because does not improve survival is incorrect, because thyroid cancer patients who are approved for PET/CT imaging represent a preselected group of patients with grave prognosis as documented by the 5-years mortality rate of 50% for patients with non-iodine avid distant metastatic disease.
6. PET imaging can only be done in thyroid cancer patients with biochemical evidence of recurrence (i.e. rising thyroglobulin levels) and negative RAI scans. PET imaging is important not only for tumor identification, - determining if disease is localized (and can be surgically resected) or distantly metastatic and/or unresectable, but also for determining the tumor metabolic signature which predicts tumor biologic behavior (as indolent tumor which can be managed with L-T4 suppression, or aggressive tumor with high likelihood of progression which requires therapeutic intervention with resection, external beam radiation, or initiation of tyrosine kinase inhibitor chemotherapy)
We request that the authors provide additional data (as above) in order to support their conclusions.
Dr. Anca M. Avram
e-mail: ancaa@umich.edu
Associate Professor of Radiology
University of Michigan Medical Center
Competing interests: No competing interests
We appreciate the letters to the Editor in response to our original research article titled, “Use of imaging tests after primary treatment of thyroid cancer in the United States: population based retrospective cohort study evaluating death and recurrence”. We agree with Dr. Pal that timing is important. Our study cohort was diagnosed between 1998-2011. The study cohort was followed through 2013 with median follow-up of 69 months. Patients diagnosed in the earlier years would have longer follow-up than those diagnosed in later years. It is possible that more diagnoses of recurrence and death could occur if follow-up were longer for the cohort diagnosed in later years. In addition, timing of imaging relative to time of treatment could differ in patients diagnosed in 1998 versus 2011 and there could also be variability in timing of imaging within diagnoses years. To clarify the incidence of recurrence, a concern of Ms. Patel, as shown in our Table 1, 6,502 patients (23% of the entire cohort) had treatment for recurrence. Depending on the cohort studied, in other studies thyroid cancer recurrence rates vary from 1-68%, with studies of diverse risk groups typically reporting a 20-30% recurrence rate.1-3 It was suggested that we compare different at risk cohorts, for example different SEER stages and different patient ages. Of note, we controlled for SEER stage and patient age when evaluating treatment for recurrence and death from thyroid cancer. It was also suggested that we evaluate younger patients. However, this cohort is from SEER-Medicare, which primarily has older patients. SEER-Medicare has the benefit of including billing data linked to a national cancer registry. SEER-Medicare is an appropriate cohort to study since death from thyroid cancer is higher in older patients. However, we understand Ms. Patel’s concern that findings may differ in younger patients and this was mentioned as a limitation in our manuscript. In summary, this manuscript is the foundation for future work. It highlights an important issue but more details are still needed to ascertain who and when to image.
References
1. Tuttle RM, Tala H, Shah J, et al. Estimating risk of recurrence in differentiated thyroid cancer after total thyroidectomy and radioactive iodine remnant ablation: using response to therapy variables to modify the initial risk estimates predicted by the new American Thyroid Association staging system. Thyroid : official journal of the American Thyroid Association 2010;20(12):1341-9.
2. Durante C, Montesano T, Torlontano M, et al. Papillary thyroid cancer: time course of recurrences during postsurgery surveillance. J Clin Endocrinol Metab 2013;98(2):636-42.
3. Grogan RH, Kaplan SP, Cao H, et al. A study of recurrence and death from papillary thyroid cancer with 27 years of median follow-up. Surgery 2013;154(6):1436-46; discussion 46-7.
Competing interests: No competing interests
To the Editor,
Medical imaging tests such as ultrasound, CT scans, and MRI are necessary to diagnose, rule out, or monitor different diseases. Although imaging tests are valuable tools, exposure to radiation has some risks. Banerjee and colleagues identified an association between imaging and treatment recurrence of differentiated thyroid cancer (1). The relationship between imaging after primary treatment and treatment recurrence suggested that there was no significant improvement in disease-specific survival, except with use of radioiodine scans in iodine avid disease. A few elements of this study provoked some questions into the previous assertion.
It would have been helpful had the author included the incidence of recurrence of thyroid cancer after primary treatment. Without this information, it is hard for the reader to understand the relative significance of recurrence and the importance of follow up imaging in the management of thyroid cancer. Lee and colleagues determined that the risk of having recurrence after initial remission of differentiated thyroid cancer is 6.7% (3).
In addition, it would have been interesting to see the authors compare the impact of imaging on treatment of recurrence in patients with regional lymph node metastasis versus those without. Patients with metastasis to regional lymph nodes have a much higher rate of recurrence and poorer outcome (4). Had the author compared these two groups, follow up imaging may have been demonstrated to be more beneficial in patient populations who are known to have better outcomes.
Age is also an important prognostic factor in the rate of recurrence of thyroid cancer. This study limited their population to patients primarily over the age of 65. It would have been helpful to assess the pattern of recurrence in patients from all age groups. According to Haymart, patients less than 20 years of age have the highest gain and greatest risks from radioactive iodine therapy (RAI). Although younger patients are more likely to respond to RAI therapy, they are at the highest risk for recurrence (2). Including this information would not only be beneficial to understanding the relationship between age and recurrence of thyroid cancer, but it can also guide future treatment therapies and research.
1. Banerjee, M., Wiebel, J. L., Guo, C., Gay, B., & Haymart, M. R. (2016, July 20). Use of imaging tests after primary treatment of thyroid cancer in the United States: Population based retrospective cohort study evaluating death and recurrence. British Medical Journal, I3839. Retrieved from http://www.bmj.com/content/354/bmj.i3839
2. Haymart, M. R. (2009, March 13). Understanding the Relationship Between Age and Thyroid Cancer. The Oncologist, 14(3), 216-221. Retrieved from http://theoncologist.alphamedpress.org/content/14/3/216.full
3. Lee, D. Y., Won, J., Choi, H. S., Park, D. J., Jung, K. C., Sung, M., Park, Y. J. (2016, July 13). Recurrence and survival after gross total removal of resectable undifferentiated or poorly differentiated thyroid carcinoma. Thyroid. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/27412715
4. Sharifi, A., Shojaeifard, A., Soroush, A., Jafari, M., Abdehgah, A. G., & Mahmoudzade, H. (2016, June 14). Predictors of Regional Lymph Node Recurrence after Initial Thyroidectomy in Patients with Thyroid Cancer. Journal of Thyroid Research, 2016, 1-5. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/27403370
Competing interests: No competing interests
Dear Editor
First of all congratulations to the author for this interesting retrospective study regarding follow up imaging in thyroid cancer.
They have compared disease specific survival rates and further treatment (surgical, radioactive iodine treatment and radiotherapy) after followup imaging with neck ultrasound, PET scan and Radioiodine scan.
However a few questions still crop up. First and foremost, the timing of follow up imaging may have a bearing on the overall disease specific and overall survival, and it needs to be seen whether it was comparable in all cases.
Second, the patients who are undergoing various modalities of treatment e.g. those undergoing surgery after neck ultrasound findings have the need of a modality to detect recurrence whether biochemical markers or an imaging study knowing the inherent nature of any malignancy to recur.
Still the study takes different view of this scenario and provides a platform for further research for better defining criteria of follow up imaging in thyroid cancer.
Competing interests: No competing interests
Extensive post op imaging need not result in survival benefit! In countries where there is no medical insurance, people are not undergoing this type of vigorous imaging and strict follow-up; this study gives them some support to go for minimal imaging such as radio isotope study only. This is very useful information.
Competing interests: No competing interests
Reply to Dr. Megan R. Haymart
Dear Dr. Haymart:
It goes without saying that authors would feel that their own “study is an important step in the right direction”, but it could be much more meaningful when this appraisal comes from your peers for the quality and significance of your work. I find it not possible to appraise your work as an “important step” when you do not specifically address critiques and questions from your peers. Instead, your brief note contains not a single answer, but partly inaccurate statements or simple truisms.
You state that “We then looked at the outcome variables[, including] recurrence and disease-specific survival to assess benefit[s] versus risks.” While you did look at disease-specific survival benefit, albeit for only 3 out of many tests (such as CT, MRI, etc.) that we do for patients with thyroid cancer, you did not weigh them against any kind of “risks”. Could you point out to readers what specific risk you assessed? The actual outcome you analyzed was re-treatment and not “recurrence”, to be exact. The clinical reality is that some patients might have been re-treated empirically, based on thyroglobulin elevation for example, which is sometimes originates from undertreated benign remnant thyroid tissue. Hence, some of your “recurrences” could very well fall into your favorite discussion category – the overdiagnosis – and not necessarily because of the over-imaging, as shown in the given example.
“Tailoring imaging to disease severity and limiting unnecessary use [of imaging] should be a goal for all physicians.” Yes, everyone would agree with this statement, which is the definition of truism that you use quite liberally. “We want the “right” patients to get the “right” imaging.” Do you know anyone who does not want the same? But how and where did this article specifically addressed any “tailoring”? One cannot just want it – you have to actually test it, if you intend to make it a point. You neither revealed information on the actual indications for the tests nor disease severity. Moreover, you included codes for brain PET/CT that are never used for thyroid cancer indication, which would be an example of the wrong imaging for the wrong patients. It served only to count PET/CT use in patients who were incidentally worked-up for dementia or seizures, which diluted the relevance of your PET/CT data. The PET/CT analysis must be considered not valid, unless you can show that the indication for other PET/CT exams was indeed for thyroid cancer, as some PET/CT exams without a doubt were done for irrelevant indications, such as evaluation of unrelated malignancy that is more common in the elderly.
This brings us to the dilemma of patients’ age handling in your study of the Medicare database, which you state has “strengths and limitations [that] were not only addressed in our paper but are also well known”. First, it may be presumptuous to think that international readership is well familiar with Medicare database. I googled looking for a recent study that used Medicare cohort in cancer research to check how others addressed the issue of younger patients. In the first paper I came across, patients younger than 65 years of age were explicitly excluded. (1) You point out that Medicare “cohort also has an increased risk of mortality from thyroid cancer due to their advanced age, which makes it the optimal cohort to identify any effect on survival.”(2) Why would you hamper this advantage by including younger patients? But regardless, a meticulous study would either clarify how many younger patients were included or exclude them altogether,(1) for the reasons you stated yourself.
You conclude that this study is “the foundation for future work and more details are needed to ascertain which patients do and do not need specific imaging”. If you read carefully my questions and criticism in this and the prior response, as well as all the questions raised by others, you should have realized that your study results are too questionable for a valid foundation. If you could do some hard work of providing the answers to each question posed to you and address thoughtfully each critique raised, it could most certainly reinforce that foundation you desire.
References:
1. Karve S, Lorenzo M, Liepa AM, Hess LM, Kaye JA, Calingaert B. Treatment Patterns, Costs, and Survival among Medicare-Enrolled Elderly Patients Diagnosed with Advanced Stage Gastric Cancer: Analysis of a Linked Population-Based Cancer Registry and Administrative Claims Database. J Gastric Cancer 2015; 15(2): 87-104.
2. Banerjee M, Wiebel JL, Guo C, Gay B, Haymart MR. Use of imaging tests after primary treatment of thyroid cancer in the United States: population based retrospective cohort study evaluating death and recurrence. BMJ 2016; 354: i3839.
Competing interests: No competing interests