We read with interest the recent article on lymph node recovery and evaluation in colorectal cancer (“In Colorectal Cancer, a Lymph Node Debate,” May 2007). There is no doubt that a sufficient number of lymph nodes must be examined to properly stage colorectal cancer, although it remains unclear when enough is enough. Recommendations typically vary between 12 and 16 nodes, but statistically rigorous breakpoints do not appear to exist; put simply, the more nodes the better. So the debate continues to rage around issues of quality, cost, and pay-for-performance as pathologists flail away in pericolic fat, hoping they will find sufficient nodes not only to provide accurate staging but also to keep themselves out of trouble. However, there is a useful cost-effective method to improve this situation—sentinel lymph node mapping, or SLNM.
Unlike the situation for melanoma or breast cancer, where SLNM is used to limit node dissection, in colorectal cancer SLNM simply identifies for the pathologist the few nodes most directly in the drainage path of the tumor, where metastasis is most likely to occur. Combined with focused examination of these nodes by multilevel sections and cytokeratin immunostaining, the technique provides a formidable tool to help accurately stage patients. While it still remains necessary to harvest additional nodes to further improve staging, the pathologist can be assured that the nodes most likely to harbor tumor have not been missed during dissection.
At our institution, this method has been used successfully for more than a decade. We recently compared SLNM cases (n=500) with conventional cases (n=368) in a multicenter trial (Saha S, et al. Am J Surg. 2006;191:305–310). For all T1–T4 tumors, 50 percent of the patients had nodal disease in the SLNM group, compared with 35 percent in the conventional group. Furthermore, in these patients followed for a minimum of two years, recurrence rates in the SLNM group were significantly lower (seven percent) than in the conventional group (25 percent).
Our findings and experience support the efficacy of the SLNM technique in colorectal cancer, and suggest that the current debate focus not only on “how many?” but on “which ones?”
David A. Wiese, PhD, MD
Department of Pathology
Sukamal Saha, MD
Department of Surgery
McLaren Regional Medical Center
Hemolysis in samples
In the May “Q&A” column, Gary L. Horowitz, MD, replies to a question about hemolyzed specimens. In that reply he recommended that potassium results be reported from grossly hemolyzed specimens. I strongly disagree. That recommendation puts patients, clinicians, and pathologists at risk, even if the laboratory appends a disclaimer about the effects of hemolysis on laboratory test results. If hemolysis occurs during the collection or handling of the specimen, then the potassium concentration of the specimen does not match the potassium concentration of the patient’s plasma. Laboratories should never report results known to be incorrect. Laboratories should not report potassium results from grossly hemolyzed specimens except for patients with intravascular hemolysis. Also, potassium results from grossly hemolyzed specimens will exceed critical result cutoffs. Except for intravascular hemolysis, hemolysis-associated high potassium results are not associated with life-threatening conditions and should not be reported as critical results.
My recommendations for hemolyzed specimens are to contact the ordering clinician or a clinician caring for the patient, explain that the specimen is hemolyzed, state which tests cannot be reported because of the hemolysis, and recommend re-collection if the clinician needs results for the unreportable tests. If the patient has intravascular hemolysis, then the laboratory should contact a clinical pathologist who will speak with the clinician and determine which test results can be released (with appropriate disclaimers).
Gregory Tetrault, MD
Department of Veterans Affairs
Gary L. Horowitz, MD, director of clinical chemistry, Beth Israel Deaconess Medical Center, Boston, and chair, CAP Chemistry Resource Committee, replies: Far from recommending that potassium results be reported from grossly hemolyzed specimens, my letter, in context, suggested that there was no single value for hemolysis above which the CAP could recommend rejecting samples and that such decisions rightly vary with conditions at each local laboratory.
I did point out that information derived from hemolyzed samples, even for potassium, could be helpful, so long as the values were reported appropriately. Furthermore, I noted that this was more than my own personal opinion, in that this position is consistent with the current CAP accreditation checklist (CHM.11900).
Dr. Tetrault feels strongly that he should not report potassium values from “grossly hemolyzed” specimens. I fully support his decision for his practice.
I encourage all laboratory directors to consider what works best for them in their own practices.
FNA and thyroid nodules
Several recommendations in the American Thyroid Association guidelines for thyroid aspiration biopsy reviewed recently in CAP TODAY (May 2007) are contrary to our experience.
We agree that thyroid nodularity is common. Medical textbooks estimate an incidence of five percent to 10 percent in the American population, currently at 300 million. This means that at least 15 million Americans have a thyroid nodule or nodules. We will never know how many have thyroid cancer, but the American Cancer Society in April predicted 33,500 cases in its 2007 assessment. In other words, a small fraction of one percent of the Americans with a nodular thyroid gland have cancer, and of that fraction less than five percent will die of the cancer. In short, thyroid nodules are common, thyroid cancer is rare, and fatal thyroid cancer is even rarer.
A second point made in the review is that non-diagnostic biopsies should be repeated with ultrasound guidance. This assumes that ultrasound guidance guarantees better results than biopsy by palpation (of palpable nodules, obviously). Our experience is that there should be very few non-diagnostic reports from aspiration biopsy with a skilled operator if the samples are assessed during the procedure. Whether done by palpation or with guidance should not be relevant.
We do not share the reported concern about cystic thyroid nodules. Benign colloid cysts are commonplace and have not been misjudged in our practice. Even hemorrhagic cysts (also common) can be assessed accurately during the procedure, though they require careful observation in the microscope to exclude a cystic hemorrhagic papillary adenocarcinoma.
Another recommendation reported in the summary of the guidelines is that patients with repeated non-diagnostic biopsies should be sent for thyroid resection. Rather, these patients should be sent to someone else for biopsy, and the original person or group should improve their practices and criteria before performing more biopsies.
Surgery is appropriate for most, but not for “all,” malignant neoplasms (think about metastatic cancers, malignant lymphomas, and anaplastic carcinomas) as recommended in the guidelines. An incidence of indeterminate diagnoses of greater than five percent is not “acceptable,” in our experience, because many benign nodules would be removed unnecessarily under the guidelines.
As recommended, suspicion of papillary adenocarcinoma must be resolved, usually at operation. We do not share the prevailing concern about oncocytic change in follicular cells, as small numbers of oncocytic cells are observed in various conditions, including nodular goiter. Fatal oncocytic follicular and papillary cancers are rare.
With regard to the various “suspicious” sonographic characteristics (small calcific deposits, solid nodules, and vascularity), these usually can be defined by a good (diagnostic) sampling and usually are benign.
Finally, the quoted false-negative rate of five percent exceeds our expectations. We should critically review falsely negative biopsies, extenuating circumstances and so-called microinvasive follicular nodules aside. However, false-negative diagnoses are equally problematic whether the biopsy is obtained by ultrasound guidance or by palpation.
The algorithm for evaluation of palpable nodules should follow the practice at the Karolinska Hospital in Sweden. Although blood tests and imaging are invaluable to understanding the results of benign aspiration biopsies, the simplest and fastest way to prove or exclude a diagnosis of cancer in a palpable nodule in the thyroid gland is to biopsy the nodule at the first opportunity.
Should we rationalize unsatisfactory results of improper aspiration biopsies or should we strive for perfection? The biopsy, preparation of samples, immediate assessment for adequacy, and interpretation in full knowledge of the patient, of the biopsy, of the specimen obtained, and of the real probabilities of the diagnosis should be made by one fully trained person with ongoing experience in thyroid problems or should be in the hands of a similarly committed, trained, and informed team.
To this end, a large group of endocrinologists, pathologists, radiologists, and surgeons will meet this fall in Bethesda to review systematically all aspects of thyroid aspiration biopsy and cytology.
Jerry Waisman, MD
Professor of Pathology
New York University
School of Medicine
New York, NY
Theodore R. Miller, MD
Professor of Pathology
University of California
San Francisco School of Medicine