Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
World J. Surg. 24, 1295–1302, 2000 DOI: 10.1007/s002680010215 WORLD Journal of SURGERY © 2000 by the Société Internationale de Chirurgie Total Thyroidectomy or Thyroid Lobectomy in Patients with Low-risk Differentiated Thyroid Cancer: Surgical Decision Analysis of a Controversy Using a Mathematical Model Electron Kebebew, M.D.,1 Quan-Yang Duh, M.D.,2 Orlo H. Clark, M.D.1 1 Department of Surgery, University of California, San Francisco, School of Medicine, UCSF/Mount Zion Medical Center, 1600 Divisadero Street, San Francisco, California 94120, USA 2 Surgical Services, Veteran Affairs Medical Center, 4200 Clement Street, San Francisco, California 94121, USA Abstract. There is a general consensus that total or near-total thyroidectomy is the optimal treatment for patients with high risk differentiated thyroid cancer (DTC), but the optimal extent of thyroidectomy in patients with low risk DTC continues to be controversial. To determine the optimal extent of thyroidectomy in patients with low risk DTC, we used decision analysis to compare the trade-offs of total thyroidectomy (TT) to thyroid lobectomy (TL). The decision analysis model included the probabilities of thyroidectomy complications, risk of DTC recurrence, and death from DTC. This information was obtained from the literature and from outcome data for patients with low risk DTC from our institution. In addition, the concept of utilities was used in the analysis. To determine the utility of each health outcome state (thyroidectomy complication, DTC recurrence, and DTC mortality for low risk patients) a survey was conducted. Overall, prospective patients viewed DTC recurrence as less desirable than thyroidectomy complication. The utilities assigned by the survey participants varied over a wide range, with 61.5% of the individuals viewing the occurrence of a thyroidectomy complication as better than DTC recurrence. At baseline utilities and probabilities, TT had a higher expected utility than TL. One-way sensitivity analysis varying the rates of (1) thyroidectomy complication, (2) DTC recurrence, and (3) DTC mortality over the possible range showed that complication from initial thyroidectomy was the most important factor that determined the preferred extent of thyroidectomy. TL was the preferred surgical approach only if a complication rate of > 33:1, TT/TL complication rate ratio, was assumed. When no differences in DTC recurrence between the two approaches was assumed in the model, TL had a higher expected utility using the baseline utilities of thyroidectomy complication and DTC mortality. The analysis indicates that TT in patients with low risk DTC is preferable to TL. However, TL is preferred if (1) no difference in the DTC recurrence rate between the two approaches is assumed, (2) a higher complication rate for TT is used (> 33 times higher), or (3) the utility ratio of thyroidectomy complication to DTC recurrence is < 0.8 TL. We believe this decision analysis model provides an objective approach that others can use to select the optimal extent of thyroidectomy based on patient preference of health outcome states, institution-specific outcome data for DTC recurrence or mortality, and the surgeon-specific complication rate. This International Association of Endocrine Surgeons (IAES) article was presented at the 38th World Congress of Surgery International Surgical Week (ISW99), Vienna, Austria, August 15–20, 1999. Correspondence to: E. Kebebew, M.D., Department of Surgery, University of California, San Francisco, 513 Parnassus Avenue, S-343, San Francisco, CA 94143-0470, USA There have been significant advances in the risk stratification of patients with differentiated thyroid cancer (papillary and follicular thyroid cancer) over the past four decades [1]. Based on clinical and pathologic data, numerous investigators have established helpful risk assessment models in patients with differentiated thyroid cancer (DTC) [2– 6]. For example, the AMES (age, distant metastasis, extrathyroid invasion, and primary tumor size) and the AGES (age, histologic grade of tumor, extrathyroid invasion and distant metastasis, and tumor size) prognostic scoring systems separate patients with DTC into high and low risk groups [3, 4]. Shaha and associates at the Memorial Sloan Kettering Cancer Center have also suggested that some patients with DTC have an intermediate risk [6]. In general, patients with large tumors, extrathyroidal tumor invasion, distant metastasis, and gross bilateral lobe tumors are considered to have high risk tumors. Although there is a general consensus that patients who have high risk DTC warrant total or near-total thyroidectomy, controversy continues as to the most optimal extent of initial thyroidectomy for most patients (about 80%) who present with low risk DTC (small, unilateral tumor) [7]. The debate has largely centered around the extent of disease, the risks associated with each surgical approach, and patient outcome (differences in DTC recurrence and mortality from DTC) (Fig. 1). A wide range of surgical approaches has been advocated for patients with DTC including thyroid lobectomy (TL), subtotal thyroidectomy, near-total thyroidectomy, and total thyroidectomy (TT) [7]. In general, proponents of TL suggest that a higher risk of recurrent laryngeal nerve injury and hypoparathyroidism are associated with the contralateral neck dissection for TT. Advocates of TT or near-TT frequently report that a lower recurrence rate and a small benefit in survival can be achieved in patients with DTC. The controversy persists largely because no prospective randomized trial evaluating the various surgical approaches for DTC has been completed. Because of the relatively low incidence of DTC and its good prognosis, such a trial would require a large, multicenter study with a long follow-up. There are, however, numerous large, long-term retrospective studies in patients with DTC that evaluated the influence of extent of initial 1296 Fig. 1. The main clinical factors considered when selecting the appropriate extent of thyroidectomy in patients with differentiated thyroid cancer have been the extent of disease and patient outcome (Cancer). Additional factors that could affect selection of the optimal surgical approach are the surgeon-specific complication rate (Surgeon) and patient preference (Patient) with respect to the health status that follows each surgical procedure. thyroidectomy on patient outcome [7]. Most studies have reported that a reduced recurrence rate and lower mortality are associated with more extensive thyroidectomy procedures [7]. Nevertheless, divergent recommendations exist for the optimal surgical resection in patients with low risk DTC even among treatment guidelines and consensus statements [8 –12]. The patient’s viewpoint and differing surgeon complication rates have not been systematically addressed in this controversy (Fig. 1). Patients express varying beliefs as to whether they would prefer to have a “higher risk” procedure if it offered a long-term lower risk of recurrence or mortality. Such factors could be important in situations where the clinical benefits of various treatment options are uncertain or debated such as exists for DTC. Decision analysis is increasingly utilized for developing patient care policies [13]. This approach is especially useful in a clinical setting where decision-making is based on uncertain or conflicting data or viewpoints [13–16]. In addition to the advantage of being able to compare clinical management approaches, decision analysis also allows incorporation of patient perspectives on health outcomes when arriving at his or her optimal treatment strategy [13, 17]. In this study, we performed a quantitative, structured risk– benefit analysis for selecting the optimal extent of thyroidectomy that incorporates patient preference and surgeon-specific complication rates. The study model considered three outcome states relative to the extent of initial thyroidectomy in patients with low risk DTC: (1) complications of thyroidectomy; (2) risk of DTC recurrence; and (3) the risk of DTC mortality (Fig. 2). Methodology Decision Model In a clinical decision analysis, a structured framework (decision tree) is used to represent the management options for identifying: (1) the treatment options; (2) consequences of the treatment options (the likely events that follow those options); and (3) the World J. Surg. Vol. 24, No. 11, November 2000 Fig. 2. Decision tree for the extent of thyroidectomy (Thx) in patients with low risk differentiated thyroid cancer. The tree includes the immediate event that follows thyroidectomy (complication) and the long-term events that occur (tumor recurrence and death from thyroid cancer). Each event is associated with a probability of a complication (Pc), recurrence (Pr), and death (Pd) and a utility (Uc, Ur). time sequence in which they occur [15, 17]. The most credible probability estimates of each of those events in the decision tree were gathered from the literature and from our experience at the UCSF/Mount Zion Medical Center over a 10-year period. The decision model used in our study is illustrated in Figure 2. It is a simple model in which two surgical approaches were compared at the time of diagnosis: thyroid lobectomy (TL) and near-total or total thyroidectomy (TT). The model includes three major, easily measured objective variables: (1) thyroidectomy complication (i.e., risk of permanent recurrent laryngeal nerve injury and permanent hypoparathyroidism); (2) DTC recurrence; and (3) death from DTC. The probability of each event from the decision node to the terminal branch is the product of the probability. Furthermore, the impact of a possible outcome on the preference of an option was weighted with the respective utility of that event occurring from the time of initial thyroidectomy. Therefore the total utility function for each surgical approach in Figure 2 is Total utility function ⫽ 共1 ⫺ Pd兲 兵共1 ⫺ Pr兲共1 ⫺ Pc兲 ⫹ UrPr共1 ⫺ Pc兲 ⫹ UcPc共1 ⫺ Pr兲 ⫾ UrPrUcPc} where P ⫽ probability of the event occurring; U ⫽ utility of health state; c ⫽ thyroidectomy complication; r ⫽ DTC recurrence; and d ⫽ death from DTC. This is the expected total utility, which is a quantitative measure of the most preferred option. The preferred extent of thyroidectomy is the approach with the highest total utility from the patient’s perspective. Outcome Variable Probabilities The analysis required establishing sound probability estimates of the thyroidectomy complication rates, the risk of DTC recurrence in low risk patients, and mortality from DTC in low risk patients. These values were determined by focused review of the available literature (in English) from long-term follow-up studies and from 10-year outcome data collected at our institution (1986 –1996). Listed in Table 1 are probability estimates for each health outcome state [18 – 47]. The baseline probability estimates were weighted averages from the listed studies based on sample size and follow-up time in patients with low risk DTC. Kebebew et al.: Thyroidectomy for Differentiated Thyroid Cancer Table 1. Probabilities for each health outcome state in the decision model culled from review of the English literature and UCSF/Mount Zion clinical outcome data in 156 patients with low risk DTC. Variable Thyroidectomy complications RLNI Hypoparathyroidism TT vs. TL UCSF TT vs. TL Baseline TT vs. TL DTC recurrence Overall TT vs. TL* UCSF TT vs. TL Baselinea DTC mortality Overall TT versus TL UCSF TT versus TL** Baseline Rate of events (%) Reported range (%) Refs. 0.077 0.05 0.39 vs. 0.075 0 vs. 0 1 vs. 2 0 –14 0 –30 [18 –38] 15.5 11.5 vs. 23.0 6.7 vs. 30.9 15.2 vs. 34.7 8.3– 40 [39 – 46] 0.1–9.0 [39 – 46] 3.2 3.2 vs. 4.6 0 vs. 5.2 3.8 vs. 6.4 RLNI: recurrent laryngeal nerve injury; DTC: differentiated thyroid cancer; TT: total thyroidectomy; TL: thyroid lobectomy; UCSF: University of California, San Francisco. a DTC recurrence at 10 to 15 years median or mean follow-up time. *Significant difference in DTC recurrence by 2 test (p ⫽ 0.0001). **Not significant (p ⫽ 0.10) by Kaplan-Meier product limit method. Utilities To determine the value individuals place on the various health outcome states, we used the concept of utilities as a quantitative measure of a decision-maker’s relative preference for an outcome state [13, 14]. This approach is based on how an individual views certain outcome states based on uncertain data balancing the risk– benefit ratio relative to the individual’s perspective. Because we did not want to use arbitrary utility estimates, we conducted a survey to obtain baseline utility estimates for each of the health outcome states. Survey participants were asked to assign a numerical value for each of the health outcome states based on anchor states of 100% (utility of 1) for a perfect state of health and 0% (utility of 0) for being dead. The survey included the following questions: (1) unilateral recurrent laryngeal nerve injury— having a hoarse voice, a weakening voice, and unable to sing high notes now and possibly for the rest of your life; (2) bilateral recurrent laryngeal nerve injury—not being able to speak and potentially requiring a tracheostomy to breath; (3) hypoparathyroidism— having to take vitamin D, calcium or both orally every day for the rest of life to avoid having muscle cramping, tingling sensation, tetany, or seizures; and (4) recurrent thyroid cancer within 10 years, which could be treated with reoperation, radioactive iodine treatment, or both. A time point for DTC recurrence of 10 years was used in the questionnaire because this length of time was most commonly reported as the mean follow-up time from which the probability estimates of recurrence and mortality were derived [39 – 46]. Furthermore, the survey questionnaire for determining the utility of recurrence and mortality from DTC accounts for the later time frame in which these events occur from initial thyroidectomy. 1297 The following assumptions were made when establishing the utility of thyroidectomy complications rates for the two surgical approaches: (1) the major difference in complication rate between the two approaches (TT versus TL) are the risk of recurrent laryngeal nerve injury and hypoparathyroidism; (2) the chance of hypoparathyroidism or bilateral laryngeal nerve injury in patients undergoing initial TL is zero because the contralateral recurrent laryngeal nerve or parathyroid gland(s) is not manipulated and thus not at risk; (3) the chance of injury to the recurrent laryngeal nerve injury is equal on both sides and independent—therefore patients who undergo TT have twice the risk of unilateral recurrent laryngeal nerve injury than patients who have TL; and (4) the total complication risk for the two approaches is the sum of the risk of hypoparathyroidism and recurrent laryngeal nerve injury (unilateral and bilateral) for each procedure. Patients with low risk DTC were defined as having TNM stage I or II DTC [47]. TT refers to the total extracapsular removal of the thyroid gland, and TL was defined as unilateral total lobectomy with isthmusectomy [7]. Decision Model Analysis The decision Markov model was constructed in a Microsoft Excel 6.0 spreadsheet. The baseline analysis was performed using the baseline probability and utility values for each health outcome state. The total expected utility was determined based on weighted products of the chance of each outcome in the decision model. The total utility function for the extent of thyroidectomy was determined such that the surgical approach yielding a higher total utility was the more desirable procedure. Sensitivity analyses were conducted by varying the utility and probabilities over the possible ranges. The threshold point was when both surgical approaches yielded an equal total expected utility (i.e., no difference existed). Results A wide range of thyroidectomy complication rates has been reported in the literature (⬍ 1% to 30%) (Table 1). In 156 patients with low risk DTC who underwent initial thyroidectomy at our institution from 1986 to 1996 (30.8% TL and 69.2% near-TT or TT), no permanent recurrent laryngeal nerve injuries or hypoparathyroidism occurred. The overall average weighted probability of thyroidectomy complication was 1% for TL and 2% for TT. In low risk patients, DTC recurrence and death from DTC have been reported to occur in 8.3% to 40.0% and 0.1% to 9.0%, respectively (Table 1). A few investigators, specifically in patients with low risk DTC, have reported DTC recurrence and mortality rates relative to the extent of thyroidectomy (Table 1) [41, 49]. At our institution, the recurrence and mortality rates from DTC in low risk patients were 6.7% and 0%, respectively, for patients who had initial TT, and 30.9% and 5.2% in patients who had initial TL, respectively (Table 1). In patients with low risk DTC, the overall average weighted probabilities of DTC recurrence were 15.2% for TT and 34.7% for TL. The lower mortality rate in low risk DTC patients who underwent TT compared to those with TL was not statistically significant in our cohort, consistent with other studies [41, 49]. The results of the questionaire indicated that, on average, individuals viewed the recurrence of DTC as less desirable than 1298 World J. Surg. Vol. 24, No. 11, November 2000 Table 2. Utilities of each health outcome state (event) in the decision treea obtained from questionnaire (n ⫽ 42). Health state Thyroidectomy complications Unilateral RLNI Bilateral RLNI Hypoparathyroidism DTC recurrence DTC mortality Dead Alive Baseline utility Range of utilities 0.627 0.205 0.778 0.54 0.10 – 0.95 0.02– 0.60 0.20 – 0.98 0.03– 0.60 0 1 — — a See Figure 2. Table 3. Rank order of possible health outcome states after initial TT and TL based on overall baseline utilities. Outcome from decision mode Compounded utilities for TT Compounded utilities for TL No complication, no recurrence, alive Complication, no recurrence, alive No complication, recurrence, alive Complication, recurrence, alive No complication, no recurrence, dead Complication, no recurrence, dead No complication, recurrence, dead Complication, recurrence, dead 1.00 0.54 0.41 0.22 0 0 0 0 1.00 0.86 0.41 0.35 0 0 0 0 having a complication from thyroidectomy (DTC recurrence/thyroidectomy complication ratio utility was 1:1.32 for TT and 1:1.53 for TL) (Table 2). The utility survey participants assigned to each outcome state, however, varied over a wide range, with 38.5% of the survey participants viewing thyroidectomy complications as less desirable than DTC recurrence. The utilities participants assigned from the questionnaire were highly reproducible (r ⫽ 0.96 by Pearson product-moment correlation cofficient), when repeated by the same individuals (n ⫽ 12). Participants assigned lower utility for TT than TL for a health outcome state of being alive with a thyroidectomy complication and a recurrence of DTC or having only a thyroidectomy complication (Table 3). At baseline probabilities, initial TT in patients with low risk DTC had a higher total expected utility (0.867) than initial TL (0.741). Based on clinical outcome data from our institution, initial TT had an even higher expected utility than TL (23.9% higher). If no difference in the probability of death from DTC was assumed between TT and TL, TT still had a higher total expected utility than initial TL. Only when a thyroidectomy complication rate of 30% for TT and 0.9% for TL was used did the total expected utility for TL become higher than that for TT. Furthermore, if no differences were assumed in the rates of DTC recurrence and DTC mortality for the two approaches, TL had a higher total expected utility at baseline probability of thyroidectomy complication. Differences in the rates of (1) thyroidectomy complication, (2) DTC recurrence, and (3) DTC mortality relative to the extent of thyroidectomy remain controversial and have inconsistently been reported as significantly different [6, 18 – 46]. Table 4 shows the various clinical outcome scenarios possible in which differences (or no differences) in the three variables can be assumed to determine the optimal extent of thyroidectomy for each scenario (i.e., TL versus near-TT or TT). Varying the probabilities of Table 4. Preferred extent of thyroidectomy in patients with low risk DTC: varying outcome variables of observed differences or no difference relative to the extent of thyroidectomy (TT vs. TL). Thyroidectomy complication rates DTC recurrence DTC mortality Preferred extent of thyroidectomy ⫹ ⫹ ⫹ ⫹ ⫺ ⫺ ⫺ ⫺ ⫹ ⫹ ⫺ ⫺ ⫹ ⫺ ⫹ ⫺ ⫹ ⫺ ⫹ ⫺ ⫹ ⫹ ⫺ ⫺ TT TT TL TL TT TL TT TT ⫹: a difference between TL and TT was used (in Pc, Pr, and Pd); ⫺: no difference between TL and TT was assumed (in Pc, Pr, and Pd). Fig. 3. Two-way sensitivity analysis of Uc versus Ur at baseline probabilities. thyroidectomy complication, DTC recurrence, or DTC mortality over the range of possible values indicated that the difference in thyroidectomy complication is the most important variable affecting the optimal extent of thyroidectomy at baseline utilities (Uc and Ur). In fact, TL becomes the preferred approach for patients with low risk DTC only when TT is expected to have more than 33 times the complication rate of TL at baseline probabilities for DTC recurrence and mortality. The finding that initial TT is preferred in patients with low risk DTC depends not only on the rates of thyroidectomy complications, DTC recurrence, and DTC mortality but also on the utilities of thyroidectomy complication (Uc) and DTC recurrence (Ur) used in the analysis. Figure 3 shows the threshold of a two-way sensitivity analysis for Uc and Ur. For example, a patient who assigns a utility of 0.5 for DTC recurrence would have to view the utility of thyroidectomy complication as ⬍ 0.4 for TL to become the preferred extent of thyroidectomy at baseline probabilities. An individual would have to view the Uc/Ur ratio as ⬍ 0.80 for TL to be the preferred extent of thyroidectomy. Discussion Surgeons who advocate TL for patients with low risk DTC cite the following reasons: (1) About half of all local DTC recurrences can Kebebew et al.: Thyroidectomy for Differentiated Thyroid Cancer be treated by reoperation, and fewer than 5% of DTC recurrences are found in the thyroid bed; (2) total thyroidectomy may be associated with a higher complication rate than lesser procedures; (3) tumor multicentricity in papillary thyroid cancer has little clinical significance; and (4) most patients with low risk DTC enjoy an excellent prognosis regardless of the extent of initial thyroidectomy [49 –54]. Surgeons who perform TT suggest the following rationale: (1) Radioiodine scanning or treatment for detecting or ablating remnant normal thyroid tissue or for local and distant metastases is more effective after all thyroid tissue is removed; (2) the postoperative serum thyroglobulin level is a more accurate marker of recurrence or persistent DTC after all normal thyroid tissue has been removed; (3) because up to 80% of papillary thyroid cancers are multicentric, possible sites of DTC recurrence in the contraleral lobe are eliminated as possible sites of recurrence; (4) although recurrent/persistent DTC may be treated with reoperation and radioiodine ablation, about 50% of these patients die from their recurrence, and thyroid reoperation may be associated with a higher complication risk; (5) about 1% of DTC progresses to anaplastic thyroid cancer, which is almost uniformly lethal; and (6) DTC recurrence is lower in patients who undergo TT than in patients who have lesser procedures, and some studies even show an improved survival rate [48 –50]. It is therefore not surprising, given these conflicting observations, that surgeons continue to debate the optimal extent of thyroidectomy. Moreover, the natural history and biologic behavior of DTC has changed over the last 30 years [49]. As previously noted, numerous prognostic scoring systems have been proposed for DTC that accurately predict patient survival and risk of DTC recurrence [1]. Unfortunately, none of these risk assessment systems relies completely on information available preoperatively. Such a system would be invaluable for prospectively planning and directing the extent of initial thyroidectomy necessary. Although a number of retrospective studies have evaluated the extent of thyroidectomy for DTC on patient outcome, as a whole the findings have been equivocal [39 – 46]. Furthermore, most of the retrospective studies that report more extensive thyroidectomy reduces the risk of DTC recurrence or mortality have not exclusively examined patients with low risk DTC [43, 44, 46]. Udelsman and colleagues have shown that a prospective randomized trial evaluating the optimal extent of thyroidectomy for patients with DTC would require an enormous study population encompassing 12,000 patients to compare thyroidectomy complication, up to 800 patients to compare DTC recurrence, and at least 3100 patients to compare DTC mortality [53]. These figures, however, were derived for a study that would include all patients with DTC. A prospective trial in patients with only low risk DTC would require an even larger study population and thus would be prohibitive. Considering these alternatives, we believe a decision analysis of the most relevant variables provides a rational, practical framework in which to study the optimal extent of the initial thyroidectomy in low risk DTC. Using the concept of utilities to measure the quality of life preference among the survey participants not only provides an estimate of the value individuals place on a health outcome state, it also allows a comparison between different possible outcomes. It is clear that patients view trade-offs between treatment differently from each other. Therefore, the baseline utilities represent an overall expected preferential outcome with interindividual variability. It is intriguing that most survey participants 1299 viewed DTC recurrence as less desirable than an immediate thyroidectomy complication. Given the survey result reproducibility, these preferences likely reflect a general public phobia of “cancer” recurrence, combined with a poor understanding of the relatively good prognosis associated with DTC. The two-way sensitivity analysis demonstrates that patients would have to view complications from thyroidectomy as 1.25 times less desirable than DTC recurrence for TL to be the preferred approach. A wide range in thyroidectomy complication rates has been reported in the literature [18 –38]. Most experts agree that the major complications of thyroidectomy are recurrent laryngeal nerve injury and hypoparathyroidism. A thyroidectomy complication depends on the extent of disease, the experience of the surgeon, the type of thyroid pathology, and the presence of any anatomic variations [7, 18, 19]. The extent of thyroidectomy has also been suggested by some to be associated with the risk of complications [49, 54]. In contrast, based on our experience and our review of studies that compared complication rates for the types of thyroidectomy, there was no significant difference between TT and lesser procedures [20, 21]. Furthermore, our analysis demonstrated that TT would have to be associated with more than 33 times the complication rate of TL for TL to yield a higher total expected utility at baseline probabilities. Most studies found that in all patients with DTC, or in patients with low risk DTC, TT was associated with a lower DTC recurrence rate [39 – 44]. For this reason and because of the lower overall utility of DTC recurrence than thyroidectomy complication, it is not surprising that our analysis suggests that among patients with low risk DTC TT had the highest total expected utility. Because patients with occult papillary thyroid cancer or minimally invasive follicular thyroid cancer essentially have near-normal life expectancy, TL would probably be sufficient in these patients [7]. Furthermore, in patients with low risk DTC who view the possibility of thyroidectomy complication as worse than DTC recurrence TL is the preferred extent of thyroidectomy. More extensive thyroidectomy has been reported to be associated with a significantly improved survival in all patients with DTC [43, 44, 46, 48]. Our experience and other studies in patients with low risk DTC show a lower mortality rate associated with TT, but the difference was not statistically significant [41, 49, 50]. This is consistent with our analysis indicating that death from DTC was not the most important factor when selecting the optimal extent of thyroidectomy. The clinical environment in which surgeons practice is rapidly changing. Patients, medical centers, and insurers are expected to have more input into the clinical decision process. Increasingly, patients also want to know about all the treatment options that are available and how it might affect their long-term quality of life. Decision analysis has been applied in many areas of medicine and is helpful for making rational management decisions based on readily available, objective clinical data [14 –17]. The quality of a decision analysis depends on a representative model being used and the accuracy of the data used to estimate outcome variables [13]. In addition to an impartial selection of outcome data from the literature, we have included our experience at UCSF/Mount Zion Medical Center. The model used in this decision analysis includes the important variables that affect thyroidectomy choices. The effects of the surgical approach relative to the utility of radioiodine scanning and serum thyroglobulin levels for follow-up of patients with DTC, however, were not included in the model 1300 and may also be important factors [7]. Furthermore, our model is limited in that it does not include a cost-effectiveness analysis [13, 15]. TL could be associated with a higher cost expenditure for the treatment of DTC recurrence and possibly a higher mortality rate. The need for lifelong thyroid hormone replacement in patients who have undergone TT may also be an important cost factor. Nonetheless, the decision analysis model used in this study is easy and practical, and it includes the most important factors used to select the extent of initial thyroidectomy for DTC. We propose that our simplified decision analysis model can be used by others to determine the optimal extent of thyroidectomy in patients with low risk DTC using outcome data from individual institutions, surgeon-specific complication rates, and patient preference. Based on our experience and data reported in the literature, TT is the treatment of choice for patients with low risk DTC. Alternatively, if a patient views the utility of the risk of thyroidectomy complication as less desirable than DTC recurrence (e.g., a professional singer) with a Uc/Ur ratio ⬍ 1.25, TL would be the preferred extent of thyroidectomy. Résumé Il y a un consensus général pour dire que la thyroı̈dectomie totale ou presque totale est le traitement optimal pour les patients ayant un risque élevé de cancer différencié de la thyroı̈de (CDT), mais le type de thyroı̈dectomie chez les patients à bas risque de CDT continue d’être sujet à controverse. Afin de déterminer le choix de l’intervention chez les patients à bas risque de CDT, nous avons comparé les résultats de risque/bénéfices (trades-off) par une analyse de la décision entre la thyroı̈dectomie totale (TT) et la lobectomie (L). Le modèle d’analyse de la décision a inclus les probabilités de complications de la thyroı̈dectomie, le risque de récidive de cancer et de décès en rapport avec le CDT. Cette information provenait de la littérature et de l’évolution des patients ayant un risque peu élevé de CDT suivis dans notre institution. De plus, on a employé le concept des utilités. On a fait une enquête pour déterminer l’«utilité» de chaque événement possible de santé (complication de la thyroı̈dectomie, récidive de cancer ou mortalité en rapport avec un cancer) pour le patient à bas risque. Globalement, les patients ont estimé que la récidive de leur cancer était «moins désirable» qu’une «complication» de la thyroı̈dectomie. Les «utilités» assignées par les participants de l’enquête ont varié dans une large gamme avec 61.5% des individus qui pensaient qu’une complication de la thyroı̈dectomie était mieux qu’une récidive d’un CDT. Avec les valeurs de base pour les «utilités» et les probabilités, la TT avaient une «utilité» attendue plus haute que la L. En variant toutes les valeurs: 1) des complications, 2) de la récidive de CDT et 3) de la mortalité du CDT, l’analyse des sensibilités a montré que les complications de la thyroı̈dectomie initiale était le facteur le plus important dans la détermination du type de la thyroı̈dectomie. La L était l’approche chirurgicale préférée seulement si on supposait que le rapport de complications TT/L était supérieur à 33. Quand on supposait qu’il n’existait pas de différences de récidive de CDT dans le modèle, la L avait une utilité attendue plus élevée compte tenue des complications éventuelles de la thyroı̈dectomie, d’une part, et de la mortalité de CDT, d’autre part. L’analyse montre que la TT est préférée à la lobectomie chez un patient à bas risque de CDT. La L, cependant, est préférée si 1) on ne suppose qu’il n’y aura aucune différence dans le taux de récidive de CDT entre les deux World J. Surg. Vol. 24, No. 11, November 2000 approches, 2) on postule que le taux de complication (avec la TT) est supérieur (plus de 33 fois), 3) le rapport des «utilités» des complications de la thyroı̈dectomie par rapport à la récidive de CDT est ⬍ 1.25. Nous croyons que le modèle de l’analyse de la décision fournit une approche objective du type de thyroı̈dectomie, basée sur les préférences des patients, sur leur état de santé, sur les chiffres des résultats spécifiques de l’institution quant à la récidive ou à la mortalité et aux taux de complications spécifiques du chirurgien. Resumen Por lo general, se acepta que el tratamiento más adecuado para pacientes de alto riesgo con cáncer diferenciado de tiroides (DTC) es la tiroidectomia total o cuasi total. Sin embargo, se discute sı́ esta extensa intervención constituye el tratamiento de elección para el DTC en pacientes de bajo riesgo. Para determinar la amplitud óptima de la tiroidectomı́a en este grupo de pacientes, se efectúa un análisis comparativo entre los resultados de la tiroidectomı́a total (TT) vs. lobectomı́a tiroidea (TL). El modelo analı́tico de toma de decisiones tuvo en cuenta las posibles complicaciones inherentes a la tiroidectomı́a, el riesgo de recidiva del DTC y la mortalidad por DTC. La información se obtuvo de la revisión bibliográfica al respecto y de los resultados de los pacientes con DTC de bajo riesgo, tratados en nuestro Hospital. En nuestro análisis se introdujo además el concepto de utilidad. Para determinar ésta se efectuó un seguimiento de cada uno de los casos, evaluando: complicaciones de la tiroidectomia, recidiva del DTC y mortalidad. En general, de forma prospectiva, los pacientes consideran que la recidiva del DTC es menos deseable que una complicación tras tiroidectomı́a. La utilidad valorada por los pacientes en seguimiento varı́a ampliamente pero un 61.5% asumen mejor una complicación tras tiroidectomı́a que una recidiva del DTC. En la lı́nea basal de utilidades y posibilidades se espera de la TT una mayor utilidad que de la TL. Un análisis lineal de la sensibilidad demuestra: 1) las diferentes complicaciones tras tiroidectomia, 2) recidiva del DTC, 3) mortalidad por DTC; el factor más importante, determinante de la amplitud de la tiroidectomı́a, fue la frecuencia de las complicaciones. La TL se prefirió sólo cuando el cociente de complicaciones entre la TT y TL fue ⬎ 33:1. Cuando la recidiva del DTC se suprimió del modelo analı́tico, la TL tuvo una mayor utilidad dada la menor tasa de complicaciones y de mortalidad por DTC. Los análisis indican que la TT es preferible a la TL en pacientes con DTC de bajo riesgo; sin embargo, la TL es la técnica de elección sı́: 1) se asume que no existe deferencia entre las dos técnicas por lo que a la recidiva se refiere, 2) cuando se registran numerosas complicaciones con la TT (superiores 33 veces a las producidas por la TL) ó 3) cuando el cociente de utilidad entre las complicaciones post-tiroidectomı́a y la recidiva del DTC es ⬍ 1.25 TL. Creemos que estos modelos analı́ticos de toma de decisiones (algoritmo) pueden constituir una aproximación objetiva para decidir la amplitud óptima de la tiroidectomı́a, pues se basa en las preferencias de los pacientes, los resultados especı́ficos de cada hospital por lo que a recidivas y mortalidad se refiere y los resultados de cada cirujano por lo que a las complicaciones atañe. Kebebew et al.: Thyroidectomy for Differentiated Thyroid Cancer 1301 Acknowledgments We thank Philip H.G. Ituarte, Ph.D., M.P.H. for his helpful suggestions and critical review of the manuscript. This work was supported in part by Mt. Zion health systems. E.K. was supported by a NIH T32 Surgical Oncology Training Grant. 22. 23. 24. References 1. Treseler, P.A., Clark, O.H.: Prognostic factors in thyroid carcinoma. Surg. Oncol. Clin. North Am. 6:555, 1997 2. Tennvall, J., Biorklund, A., Moller, T., Ranstom, J., Akerman, M.: Is the EORTC prognostic index of thyroid cancer valid in differentiated thyroid carcinoma? Cancer 57:1405, 1986 3. Cady, B., Rossi, R.: An expanded view of risk-group definition in differentiated thyroid carcinoma. Surgery 104:947, 1988 4. Hay, I.D., Grant, C.S., Taylor, W.F., McConahey, W.M.: Ipsilateral lobectomy versus bilateral lobar resection in papillary thyroid carcinoma: a retrospective analysis of surgical outcome using a novel prognostic scoring system. Surgery 102:1088, 1987 5. Hay, I.D., Bergstralh, E.J., Goellner, J.R., Ebersold, J.R., Grant, C.S.: Predicting outcome in papillary thyroid carcinoma: development of a reliable prognostic scoring system in a cohort of 1779 patients surgically treated at one institution during 1940 through 1989. Surgery 114:1050, 1993 6. Shaha, A.R., Loree, T.R., Shah, J.P.: Prognostic factors and risk group analysis in follicular carcinoma of the thyroid. Surgery 118:1131, 1995 7. Kebebew, E., Clark, O.H.: Differentiated thyroid cancer: “complete” rational approach. World J. Surg., 24:942, 2000 8. Singer, P.A., Cooper, D.S., Daniels, G.H., Ladenson, P.W., Greenspan, F.S., Levy, E.G., Braverman, L.E., Clark, O.H., McDougall, I.R., Ain, K.V., Dorfman, S.G.: Treatment guidelines for patients with thyroid nodules and well-differentiated thyroid cancer. Arch. Intern. Med. 156:2165, 1996 9. American Association of Clinical Endocrinologists clinical practice guidelines for the management of thyroid carcinoma. Endocr. Pract. 3:60, 1997 10. Shaha, A.R., Byers, R.M., Terz, J.J.: Thyroid cancer surgical practice guidelines: scope and format of guidelines. Oncology 11:1228, 1997 11. Pasieka, J.L., Rotstein, L.E.: Consensus conference on well-differentiated thyroid cancer: a summary. Can. J. Surg. 36:298, 1993 12. Clark, O.H., Levin, K., Zeng, Q.H., Greenspan, F.S., Siperstein, A.E.: Thyroid cancer: the case for total thyroidectomy. Eur. J. Cancer Clin. Oncol. 24:305, 1988 13. Petitti, D.B.: Meta-analysis, decision analysis, and cost-effectiveness analysis: methods for quantitative synthesis in medicine. Monogr. Epidemiol. Biostat. 24:20, 1994 14. Powe, N.R., Danese, M.D., Ladenson, P.W.: Decision analysis in endocrinology and metabolism. Endocrinol. Metab. Clin. North Am. 26:89, 1997 15. Danese, M.D., Powe, N.R., Sawin, C.T., Ladenson, P.W.: Screening for mild thyroid failure at the periodic exam: a decision and costeffectiveness analysis. J.A.M.A. 276:285, 1996 16. Friedman, L.S., Roberts, M.S., Brett, A.S., Marton, K.I.: Management of asymptomatic gallstones in the diabetic patient: a decision analysis. Ann. Intern. Med. 109:913, 1988 17. Olak, J., Detsky, A.: Surgical decision analysis: esophagectomy/ esophagogastrectomy with or without drainage? Ann. Thorac. Surg. 53:493, 1992 18. Sosa, J.A., Bowman, H.M., Tielsch, J.M., Powe, N.R., Gordon, T.A., Udelsman, R.: The importance of surgeon experience for clinical and economic outcomes from thyroidectomy. Ann. Surg. 228:320, 1998 19. Van Zuidewijn, W.D.R., Songun, I., Kievit, J., van de Velde, J.H.: Complications of thyroid surgery. Ann. Surg. Oncol. 2:56, 1995 20. Moulton-Barrett, R., Crumley, R., Jalilie, S., Segina, D., Allison, G., Marshak, D., Chan, E.: Complications of thyroid surgery. Int. Surg. 82:63, 1997 21. Reeve, T.S., Curtin, A., Fingleton, L., Kennedy, P., Mackie, W., Porter, T., Simons, D., Townend, D., Delbridge, L.: Can total thyroidectomy be performed as safely by general surgeons in provincial 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. centers as by surgeons in specialized endocrine surgical units? Arch. Surg. 129:834, 1994 Sherman, L.J., Strong, E.W.: Complications after total thyroidectomy. Otolaryngol. Head Neck Surg. 101:472, 1989 Shindo, M.L., Sinha, U.K., Rice, D.H.: Safety of thyroidectomy in residency: a review of 186 consecutive cases. Laryngoscope 105:1173, 1995 Reeve, T.S., Delbridge, L., Cohen, A., Crummer, P.: Total thyroidectomy: the preferred option for multinodular goiter. Ann. Surg. 206: 782, 1987 Liu, Q., Djuricin, G., Prinz, R.A.: Total thyroidectomy for benign thyroid disease. Surgery 123:2, 1998 Hines, J.R., Winchester, D.J.: Total lobectomy and total thyroidectomy in the management of thyroid lesions. Arch. Surg. 128:1060, 1993 Schroder, D.M., Chambors, A., France, C.J.: Operative strategy for thyroid cancer: is total thyroidectomy worth the price? Cancer 58: 2320, 1986 Flynn, M.B., Lyons, K.J., Tarter, J.W., Ragsdale, T.L.: Local complications after surgical resection for thyroid carcinoma. Am. J. Surg. 168:404, 1994 Harness, J.K., Fung, L., Thompson, N.W., Burney, R.E., McLeod, M.K.: Total thyroidectomy: complications and technique. World J. Surg. 10:781, 1986 Clark, O.H.: Total thyroidectomy: the treatment of choice for patients with differentiated thyroid cancer. Ann. Surg. 196:361, 1982 Deaconson, T.F., Wilson, S.D., Cerletty, J.M., Komorowski, R.A.: Total or near total thyroidectomy versus limited resection for radiation-associated thyroid nodules: a twelve-year follow-up of patients in a thyroid screening program. Surgery 100:1116, 1986 Pederson, W.C., Johnson, C.L., Gaskill, H.V., Aust, J.B., Cruz, A.B.: Operative management of thyroid disease: technical considerations in a residency training program. Am. J. Surg. 148:350, 1984 Attie, J.N., Moskowitz, G.W., Margouleff, D., Levy, L.M.: Feasibility of total thyroidectomy in the treatment of thyroid carcinoma: postoperative radioactive iodine evaluation of 140 cases. Am. J. Surg. 138: 555, 1979 Perzik, S.L.: The place of total thyroidectomy in the management of 909 patients with thyroid disease. Am. J. Surg. 132:480, 1976 Schwartz, A.E., Freidman, E.W.: Preservation of the parathyroid glands in total thyroidectomy. Surg. Gynecol. Obstet. 165:327, 1987 Martensson, H., Terins, J.: Recurrent laryngeal nerve palsy in thyroid gland surgery related to operations and nerves at risk. Arch. Surg. 120:475, 1985 Ley, P.B., Roberts, J.W., Symmonds, R.E., Hendricks, J.C., Snyder, S.K., Frazee, R.C., Smith, R.W., McKenney, J.F., Brindley, G.V.: Safety and efficacy of total thyroidectomy for differentiated thyroid carcinoma: a 20 year review. Am. Surg. 59:110, 1993 Foster, R.S.: Morbidity and mortality after thyroidectomy. Surg. Gynecol. Obstet. 146:423, 1978 Rossi, R.L., Cady, B., Silverman, M.L., Wool, M.S., Horner, T.A.: Current results of conservative surgery for differentiated thyroid carcinoma. World J. Surg. 10:612, 1986 Kokkonen, S.T., Haapiainen, R.K., Franssila, K.O., Sivula, A.H.: Papillary thyroid carcinoma: the new, age-related TNM classification system in a retrospective analysis of 199 patients. World J. Surg. 14:837, 1990 Hay, I.D., Grant, C.S., Bergstralh, E.J., Thompson, G.B., van Heerden, J.A., Goellner, J.R.: Unilateral total lobectomy: is it sufficient surgical treatment for patients with AMES low-risk papillary thyroid carcinoma? Surgery 124:958, 1998 Cady, B.: Our AMES is true: how an old concept still hits the mark: or, risk group assignment points the arrow to rational therapy selection in differentiated thyroid cancer. Am. J. Surg. 174:462, 1997 DeGroot, L.J., Kaplan, E.L., McCormick, M., Straus, F.H.: Natural history, treatment, and course of papillary thyroid carcinoma. J. Clin. Endocrinol. Metab. 71:414, 1990 Loh, K.C., Greenspan, F.S., Gee, L., Miller, T.R., Yeo, P.P.B.: Pathological tumor-node-metastasis (pTNM) staging for papillary and follicular thyroid carcinomas: a retrospective analysis of 700 patients. J. Clin. Endocrinol. Metab. 82:3553, 1997 1302 45. Grant, C.S., Hay, I.D., Gough, I.R., Bergstralh, E.J., Goellner, J.R., McConahey, W.M.: Local recurrence in papillary thyroid carcinoma: is extent of surgical resection important? Surgery 104:954, 1988 46. Samaan, N.A., Maheshwari, Y.K., Nader, S., Hill, C.S., Schultz, P.N., Haynie, T.P., Hickey, R.C., Clark, R.L., Goepfert, H., Ibanez, M.L., Litton, C.E.: Impact of therapy for differentiated carcinoma of the thyroid: an analysis of 706 cases. J. Clin. Endocrinol. Metab. 56:1131, 1983 47. Hay, I.D.: Papillary thyroid carcinoma. Endocrinol. Metab. Clin. North Am. 19:545, 1990 48. Mazzaferri, E.L., Jhiang, S.M.: Long-term impact of initial surgical and medical therapy on papillary and follicular thyroid cancer. Am. J. Med. 97:418, 1994 49. Shaha, A.R., Shah, J.P., Loree, T.R.: Low-risk differentiated thyroid cancer: the need for selective treatment. Ann. Surg. Oncol. 4:328, 1997 World J. Surg. Vol. 24, No. 11, November 2000 50. Shah, J.P., Loree, T.R., Dharker, D., Strong, E.W.: Lobectomy versus total thyroidectomy for differentiated carcinoma of the thyroid: a matched-pair analysis. Am. J. Surg. 166:331, 1993 51. Jossart, G., Clark, O.H.: Well differentiated thyroid cancer. Curr. Probl. Surg. 31:938, 1994 52. Chen, H., Udelsman, R.: Papillary thyroid carcinoma: justification for total thyroidectomy and management of lymph node metastases. Surg. Oncol. Clin. North Am. 7:645, 1998 53. Udelsman, R., Lakatos, E., Ladenson, P.: Optimal surgery for papillary thyroid carcinoma. World J. Surg. 20:88, 1996 54. Cohn, K.H., Backdahl, M., Forsslund, G., Auer, G., Zetterberg, A., Lundell, G., Granberg, P-O., Lowhagen, T., Willems, J-S., Cady, B.: Biologic considerations and operative strategy in papillary thyroid carcinoma: arguments against the routine performance of total thyroidectomy. Surgery 96:957, 1984