This article is more than 5 years old. Information may no longer be current.
Symposium findings provide ‘wake-up call’ about postoperative complications, registry discrepancies
Click Here to Manage Email Alerts
The Society of Surgical Oncology’s 69th Annual Cancer Symposium, held March 2-5 in Boston, featured the theme “Tomorrow’s Treatments for Today’s Patients.”
The meeting — which attracted approximately 1,800 surgical oncologists and other health care professionals — featured more than 500 abstracts that offered insights into therapeutic advances for patients with solid tumors.
I will review in detail three presentations I found particularly compelling.
Cost of PET/CT for liver metastases
In an era in which quality of patient cancer care, outcomes and cost are rightly being emphasized, one of the more interesting presentations came from Pablo Serrano Aybar, MD, of McMaster University in Canada.
Serrano Aybar and colleagues evaluated PET/CT scan vs. no PET/CT scan in the management of potentially resectable colorectal cancer liver metastases.
The presentation centered on the cost implications of a previous randomized controlled trial conducted by Moulton and colleagues, results of which were published in 2014 in JAMA. The multi-institutional randomized trial included 400 patients with colorectal cancer; of these, 270 underwent a PET/CT scan and 130 served as controls.
This trial showed that, although treatment plans changed for 21 (8%) of the 270 patients who underwent a PET/CT scan, the percentages of patients in the PET/CT group and the control group who underwent hepatic resection were nearly identical (91% vs. 92%). More importantly, after a follow-up period of up to 3 years, results showed no statistically significant difference in OS overall between groups, or among those who underwent surgical resection vs. those who did not.
Serrano Aybar and colleagues conducted a cost analysis of the PET/CT vs. no PET/CT regimens up to 1 year following randomization of the clinical trial.
They used administrative databases to obtain unit costs for hospitalization, physician services, chemotherapy and outpatient radiology and endoscopic procedures. The researchers undertook cost analysis from the perspective of a third-party payer, and they used a Bayesian approach to estimate mean cost in Canadian dollars.
The results showed the estimated mean cost per patient was CAN $45,454 in the PET/CT arm vs. CAN $40,859 in the control arm, equating to a difference of CAN $4,327 (95% CI, –2,207 to 10,614).
Interestingly, the primary cost driver was the cost of hospitalization for hepatic surgery (CAN $2,997 for the PET/CT arm), primarily due to a longer length of hospital stay for the PET/CT arm than the control (median, 7 days vs. 6 days; P = .0034) and a higher rate of postoperative complications (20% vs. 10%; P = .0014).
Baseline characteristics appeared similar between groups, including a similar number of liver segments involved with cancer, the number of segments resected and the type of hepatic resection performed. However, the researchers speculated the higher rate of complications in the PET/CT arm may have been secondary to a higher number of abnormal liver segments.
Serrano Aybar and colleagues concluded in this very provocative cost analysis that PET/CT does not appear to provide a significant clinical benefit in the surgical management of patients with resectable colorectal cancer liver metastases and does not save money compared with the control regimen.
It is certainly important that cost analysis in the surgical oncology world has come to the forefront, and the researchers ought to be congratulated on this analysis. Most of my surgical oncology colleagues would agree there is definitely imaging overkill in the evaluation of patients with a diagnosis of cancer.
Of course, this study was done in Canada, and one of the questions that arises is: Can the results be reproducible in the United States? The other question is whether it is worth investing the time, effort and resources to try to reproduce these results. The readers can make their own choice.
Registry discrepancies
Colleen Kiernan, MD, of the department of general surgery at Vanderbilt University, offered an interesting presentation based on a retrospective review of the Tennessee Cancer Registry.
Interestingly, studies that use the National Cancer Data Base show that 20% of patients with thyroid lobectomy also receive radioactive iodine, even though radioactive iodine after thyroid lobectomy is nonstandard care.
Kiernan and colleagues hypothesized that a considerable number of thyroid cancer registry abstracts have the variable surgery of the primary site coded inaccurately.
The researchers reviewed the thyroid cancer database of the Tennessee Cancer Registry, which obtains case information from health care providers and facilities that diagnose or treat cancer patients. The registry classifies hospitals based on whether they are accredited by the Commission on Cancer.
Certified registrars within the Tennessee Cancer Registry reviewed the abstracted text or called the reporting facility staff to verify that the definitive procedure used indeed was thyroid lobectomy.
The investigators also reviewed a subgroup of records that were initially coded with thyroid lobectomy as the definitive procedure and indicated postoperative receipt of radioactive iodine.
The analysis included 918 thyroid cancer cases diagnosed or treated at Tennessee facilities from 2004 to 2011 that were coded with thyroid lobectomy. Of these, 369 (40.2%) were incorrectly coded.
A majority (n = 242; 65.6%) of the incorrectly coded cases were changed from thyroid lobectomy to total thyroidectomy. Of the 242 cases that were changed, 85% were reported from facilities accredited by the Commission on Cancer.
Results showed 184 (20%) abstracts originally were coded with thyroid lobectomy and reported receipt of postoperative radioactive iodine. Of these, 115 (62.5%) were incorrectly coded.
Hence, after the review, only 7.5% of thyroid lobectomy cases received radioactive iodine (P < 0.01 for before-and-after comparison).
Thyroid cancer registries that include the extent of surgical procedure are at risk for inaccurate coding, Kiernan and colleagues concluded.
In this one state cancer registry, approximately 26% of cases originally coded as thyroid lobectomy should have been coded as total thyroidectomy. More importantly, most of the incorrectly coded cases were reported from facilities accredited by the Commission on Cancer.
One must keep in mind that these Commission on Cancer facilities contribute case information to other large national databases, such as the National Cancer Data Base. This provocative study demonstrates that these discrepancies in cancer registries need to be corrected.
Kiernan and colleagues suggested that use of text-to-code re-abstraction audits, along with facility contact when required, could help reduce these discrepancies.
Complications after complex cancer resection
Hari Nathan, MD, PhD, assistant professor of surgery at University of Michigan, presented an abstract titled “Postoperative complications and long-term survival after complex cancer resection.”
This presentation was based on recent information that focused on the ability to rescue patients from postoperative complications and prevent short-term mortality. However, few studies in the literature examined whether rescued patients have similar long-term outcomes as those with uncomplicated cancer resections.
Nathan and colleagues used SEER–Medicare data from 2005 to 2009 to evaluate patients aged 65 years or older who underwent resection for cancers of the lung, esophagus or pancreas.
The researchers defined serious complications as those associated with length of stay longer than the 75th percentile. They used Cox proportional hazards models to adjust survival for age, sex, comorbidity, resection subtype and disease stage.
The analysis included 15,326 cases, of which 12,395 (80.8%) were lung resection, 1,966 (12.8%) were pancreas resection and 965 (6.2%) were esophageal resection.
Researchers reported 30-day mortality rates of 6.2% in the esophageal resection group, 3.9% in the pancreas resection group and 3.3% in the lung resection group. They reported serious complications among 17% of the esophageal resection group, 12% of the pancreas resection group and 10% of the lung resection group.
Patients who developed serious complications appeared less likely than those with no complications to survive 5 years (20% vs. 43% for esophageal resection; 10% vs. 21% for pancreas resection; and 29% vs. 54% for lung resection). This trend persisted even among patients who were rescued and survived 30 days.
The investigators even reported shorter risk-adjusted long-term survival among patients who developed serious complications but survived 180 days after surgery.
Infectious complications accounted for nearly half (47%) of all complications developed by those who underwent esophageal resection, compared with 45% among those who underwent pancreas resection and 22% who underwent lung resection.
The development of infectious complications appeared associated with reduced 5-year survival among patients who underwent lung resection (26% vs. 32%; P < .0001) and pancreas resection (9% vs. 15%; P < .01).
Analyses that accounted for age and disease stage showed patients who developed serious complications were significantly less likely to receive chemotherapy. For example, among patients who underwent resection for stage II lung cancer, 53% of those who developed no complications received chemotherapy compared with 19% of those who developed serious complications (P < .0001).
Patients who undergo complex cancer resection and experience serious complications have diminished long-term survival even if they are rescued and survive more than 6 months after surgery, Nathan and colleagues concluded.
Despite the known limitations of the SEER database, the findings from this study should be a wake-up call that the 30-day cutoff for postoperative complications is too short a time to evaluate the true consequences of surgical complications.
References:
Kiernan CM, et al. Abstract 5.
Nathan H, et al. Abstract 81.
Serrano Aybar P, et al. Abstract 9.
Moulton CA, et al. JAMA. 2014;doi:10.1001/jama.2014.3740.
For more information:
Nicholas J. Petrelli, MD, is medical director at Helen F. Graham Cancer Center and Research Institute. He also is HemOnc Today’s Associate Editor for Surgical Oncology. He can be reached at Christiana Care Health System, 4701 Ogletown Stanton Road, Newark, DE19713; email: npetrelli@christianacare.org.
Disclosure: Petrelli reports no relevant financial disclosures.