HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Personal Folders Download to citation manager Author home page(s): Joseph B. Shrager Zhen Chen Larry R. Kaiser Permission Requests Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Shrager, J. B. Articles by Kaiser, L. R. Search for Related Content PubMed PubMed Citation Articles by Shrager, J. B. Articles by Kaiser, L. R. Related Collections Mediastinum

Ann Thorac Surg 2006;82:1863-1869
© 2006 The Society of Thoracic Surgeons

---

Original Articles: General Thoracic

Outcomes After 151 Extended Transcervical Thymectomies for Myasthenia Gravis

^a Department of Surgery, Division of Thoracic Surgery, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania
^b Department of Epidemiology and Biostatistics, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania

Accepted for publication May 22, 2006.

^_* Address correspondence to Dr Shrager, University of Pennsylvania, 4 Silverstein Bldg, HUP, 3400 Spruce St, Philadelphia, PA 19104 (Email: joseph.shrager{at}uphs.upenn.edu).

Presented at the Poster Session of the Forty-second Annual Meeting of The Society of Thoracic Surgeons, Chicago, IL, Jan 30–Feb 1, 2006.

General thoracic surgery: The Annals of Thoracic Surgery CME Program is located online at http://cme.ctsnetjournals.org. To take the CME activity related to this article, you must have either an STS member or an individual non-member subscription to the journal.

 

	Abstract

Top Abstract Introduction Patients and Methods Results Comment Acknowledgments References

 
BACKGROUND: The ideal operative technique for thymectomy in myasthenia gravis (MG) remains controversial. We present the largest series of extended transcervical thymectomy to provide outcomes data to compare with transsternal procedures.

METHODS: A retrospective chart review/interview was made of 164 patients operated upon from 1992 to 2004. Complete remission (CR) was defined as asymptomatic off medication for 6 months or asymptomatic on low-dose single-drug therapy ( 10 mg/d prednisone or 150 mg/d azathioprine). A modified Osserman classification based upon the Myasthenia Gravis Foundation of America quantitative disease severity score was employed.

RESULTS: The overall complication rate was 7.3%, and nearly all procedures were outpatient. Mean age at surgery was 43 years, and mean preoperative Osserman class was 2.3 (21% class 1; 39% class 2; 28% class 3; 12% class 4). Mean length of follow-up was 53 months. Mean postoperative Osserman class was 1.0. Nineteen percent of patients failed to improve. The crude cumulative CR rate was 37% (n = 58). Kaplan-Meier estimates of CR were 43% and 45% at 3 and 6 years, respectively. On multivariate analysis, only preoperative disease severity was significantly (inversely) associated with Kaplan-Meier CR rates. Longer-term follow-up (83 months) of only the earlier patients shows preserved CR rates (46%).

CONCLUSIONS: This largest series of extended transcervical thymectomy for MG confirms that the 5-year Kaplan-Meier CR rate is comparable with that obtained after transsternal procedures. Patients with less severe disease have higher CR rates. Complete responses are durable, as the CR rate remains stable with extended follow-up.

Although its value has never been evaluated in a randomized trial, and therefore some controversy continues over its use, thymectomy as a treatment strategy in myasthenia gravis (MG) is supported by a preoponderance of retrospective data. Improvement in myasthenic symptoms after thymectomy occurs in 60% to 100% of patients [1], and drug-free complete remission occurs in 19% to 50% of patients [2, 3]. These results appear to offer a clear improvement over historical controls who did not receive thymectomy.

The operative approach and extent of dissection for thymectomy in MG has also remained controversial since its initial description in 1912 [4]. Most surgeons perform an operation through median sternotomy that removes primarily the encapsulated thymus with varying extents of adjacent mediastinal fat. Some, on the basis of, first, the common presence of widely distributed ectopic thymic tissue [5], and second, reports of complete responses to excision of residual thymic tissue after failed initial thymectomy, have advocated a more extensive cervical and transsternal mediastinal dissection designed to reliably extirpate all sites of thymic tissue. Others, including our group [6, 7], have taken the approach of removing only the encapsulated gland and adjacent fatty tissue but have been able to do this through a transcervical approach that dramatically reduces postoperative morbidity and disability.

There is no doubt that the largely outpatient, extended transcervical thymectomy (TCT) procedure is less morbid and less costly than transsternal procedures which result in major complication rates between 4% [8] and 22% [9]. What remains controversial is whether response rates are similar. We previously reported a Kaplan-Meier complete response rate of 43% at 5 years after TCT [7], a value that compares favorably with results after transsternal procedures. We now report data on a larger group of patients—representing the largest accumulated series of extended TCT—to determine whether our initial findings are confirmed. In addition, we examine data from the earlier patients in our series (now with the benefit of longer follow-up) to determine if response rates after TCT remain stable over time.

	Patients and Methods

Top Abstract Introduction Patients and Methods Results Comment Acknowledgments References

 
We conducted a retrospective review of 164 consecutive patients who underwent attempted TCT for MG between January 1992 and December 2004. The Institutional Review Board of the University of Pennsylvania approved the study and granted a waiver of the requirement for informed consent. The patients' medical records, operative notes, and pathology reports were reviewed, and patients were contacted when necessary to obtain updated follow-up. Of the 164 patients, 13 (7.9%) required extension of the procedure beyond the cervical incision (2 thoracoscopies, 11 partial or full sternotomies). Of the remaining 151, 147 (97.4%) were available for follow-up by either telephone contact or communication with their neurologist and are included in the response rate analyses. All 151 patients are included in the description of all other results.

The patients' disease was categorized by a modified Osserman classification: 1, ocular signs and symptoms; 2, mild generalized weakness; 3, moderate generalized weakness; 4, severe generalized weakness, respiratory dysfunction, or both. Because this is the classification that was employed in most patients throughout their preoperative care and evaluations as well as in our previous publication on this topic, we considered it was most appropriate to continue to use this classification in continued postoperative follow-up. However, in the telephone script used to obtain follow-up data, questions were based upon the more detailed Myasthenia Gravis Foundation of America (MGFA) clinical classification and quantitative MG score for disease severity [10]. Answers to these questions were then used to fit the patients into one of the modified Osserman categories. The preoperative severity of disease noted is the maximum preoperative severity the disease reached.

The procedure performed consisted of an extended TCT, as described by Cooper and associates [11], involving extracapsular removal of the entire gland including the cervical and mediastinal poles as well as the bulk of the extrathymic mediastinal fat between the phrenic nerves and down to the diaphragm. The procedure does not remove the pleurae or fat directly apposed to the pleurae, tissue posterior to the phrenic nerves, or other areas where ectopic thymic tissue has been described. No drains are left in place. The occasional MG patient who is unable to completely extend his or her neck, and patients with suspected thymomas that appear invasive or are greater than 3 cm in diameter, are operated upon through median sternotomy and are not included in this analysis.

Although in our previous work we defined a complete response (CR) as asymptomatic off all medications for at least 6 months, we have recently become aware of a practice of our neurologist that renders this definition misleading. We have learned that despite absence of symptoms for as long as several years after thymectomy, our neurologists are hesitant to completely stop immunosuppressive agents because of reports in the literature of relapses occurring in this setting [12]. Given the minimal morbidity of maintaining a patient on a single-drug regimen consisting of 5 to 10 mg prednisone daily or 100 to 150 mg azathioprine daily, several of our neurologists will maintain patients on one of these drugs indefinitely even in the absence of symptoms. Because of this practice, we have redefined CR for this study as either asymptomatic off all medications for at least 6 months, or asymptomatic and on single-drug, low-dose immunosuppressive therapy ( 10 mg/d prednisone or 150 mg/d azathioprine) for at least 6 months. Our primary analyses define CR in this manner. In some areas, we also report results using the prior, stricter definition of a CR for the sake of completeness, and we indicate in the text where this is the case.

Crude cumulative CR rate (number of CRs divided by number of patients evaluated) was calculated. More appropriately, the Kaplan-Meier estimate of time to complete remission and CR rates at specific time points were also determined. For Kaplan-Meier estimation [13], time to CR was defined as time from surgery to first date that CR was achieved (start date). These patients then had to remain a CR for at least 6 months' follow-up time from the start date to be counted as a CR. Those who had not achieved a CR were censored, and their time to CR was defined as time from surgery to most recent patient contact. These non-CR patients had to have at least 6 months of follow-up after surgery to be included in the analysis.

Univariate survival analysis using the log-rank test [14] was performed to determine the impact of various preoperative and operative factors upon time to CR. The Cox proportional hazards regression model [15] was employed to determine the simultaneous impact of several prognostic factors. Statistical significance was set at the 0.05 level. All analyses were performed in R language and environment 2.2.0 (R Development Core Team, 2005).

	Results

Top Abstract Introduction Patients and Methods Results Comment Acknowledgments References

 
Patient Characteristics
The mean age of the 151 patients at surgery was 42.5 years (range, 13 to 78), and 60.3% were female. Mean preoperative Osserman class was 2.3: 21.2% class 1, 39.1% class 2, 27.6% class 3, and 12.2% class 3. Preoperative duration of symptoms was greater than 24 months in 31.4% of those for whom this information was available and less than 24 months in 68.6%.

Preoperative medical treatment regimens consisted of single-drug therapy in 75 patients (62 pyridostigmine [PYR], 11 steroid [ST], 1 azathioprine [AZ], and 1 Cellcept [CC]). Two-drug or modality therapy was used in 51 patients (most commonly PYR/ST [n = 21] or PYR/plasmapheresis [n = 14]). Three-drug or modality therapy was used in 17 patients (most commonly PYR/ST/plasmapheresis [n = 5] or PYR/ST/AZ [n = 4]). Four-drug or modality therapy was used in 7 patients (most commonly PYR/ST/AZ/plasmapheresis [n = 4]). One patient was receiving no medical therapy at the time of her surgery.

Operative Data and Morbidity
In our initial group of 78 patients, mean operative time (incision to closure) was 96.8 minutes; we did not measure this variable in our subsequent patients. Mean hospital stay in all 151 TCT patients was 1.10 days (using 0.5 days = outpatient). In our last 73 patients, mean hospital stay was 1.01 days. Of these, all but one single operation were performed on an outpatient basis. There was no perioperative mortality. There were 2 late deaths (3 and 3.7 years postoperatively). Table 1 shows that only 11 patients (7.3%) suffered any operative morbidity, and of these, only 1 (0.66%) was considered a major morbidity (vocal cord paralysis).

Crude Cumulative CR Rates
Table 2 shows the crude cumulative response rates to extended TCT. Patients were considered "improved" if either they improved by at least one Osserman class and were on the same amount or less medication, or if they were in the same Osserman class but on less medication. By this definition, 80.8% of patients were improved. Mean postoperative Osserman class was 1.0.

View larger version (8K): [in this window] [in a new window]   Fig 1. Kaplan-Meier curve for time to complete remission (CR) for all transcervical thymectomy patients. (MG = myasthenia gravis.)

View larger version (8K): [in this window] [in a new window]   Fig 2. Kaplan-Meier curve for time to complete remission (CR) for the 84 patients operated upon before September 1999, now with extended follow-up. (MG = myasthenia gravis.)

Univariate Analyses
Table 3 lists the factors evaluated by univariate analysis and the results of that analysis. Note that the only factor found to be significantly associated with the Kaplan-Meier CR rate at 5 years was preoperative severity of disease as reflected in the modified Osserman class. A lower Osserman class predicted a higher Kaplan-Meier CR rate (p = 0.03; Fig 3).

View larger version (12K): [in this window] [in a new window]   Fig 3. Kaplan-Meier curve for time to complete remission (CR) according to Osserman class. (Solid line = vision symptoms [n1]; dashed line = mild symptoms [n2]; dotted line = moderate symptoms [n3]; dot-dashed line = severe symptoms [n4].)

Multivariate Analysis
A multivariate Cox regression analysis was performed using only Osserman class and histology as the adjusting factors. These were selected because aside from preoperative drug therapies, they alone have both support in the literature as significant prognostic factors and p values on univariate analysis less than 0.4. We did not use preoperative treatment modalities in the Cox model because we felt that they were likely merely additional markers of severity of disease.

With this multivariate analysis, preoperative Osserman Class remained a significant factor while histology remained insignificant (Table 4).

	Comment

Top Abstract Introduction Patients and Methods Results Comment Acknowledgments References

However, we believe that a practice pattern by our neurologists that we did not previously recognize renders our altered definition of a CR—to include patients who are asymptomatic but taking either 10 mg/d or less prednisone or 150 mg/d or less azathioprine—a fairer reflection of true complete response than our prior definition of CR. We have discovered that our neurologists often maintain patients who have been asymptomatic for years on low doses of one of these immunosuppressive medications. Since many patients remain in this category (asymptomatic but on a low-dose immunosuppressive drug) for prolonged periods while additional patients will be added to this group over time, the proportion of the entire cohort that this group represents will increase over time. Thus, in this complete series, even with actual results that were the same as in our earlier series, with the passing of 5 more years there would be a larger proportion of patients in the asymptomatic and low-dose immunosuppressive therapy group and a lower proportion in drug-free complete remission. Because many of the patients in the asymptomatic or low-dose immunosuppression therapy group would likely remain asymptomatic off those medications, we believe that they are appropriately included within the definition of a CR. Including these patients, the data show that the CR rate for our entire series is almost identical to what it had been at the interim analysis, at 45%.

Since publication of our earlier series, two additional important papers have appeared in the literature addressing results after thymectomy for MG, to which our results should be compared. These include the studies by Huang and colleagues [16] and dePerrot and associates [17]. The Huang study presents the highest crude cumulative CR rate reported in recent years after a transsternal approach—58% with 8.5 year mean follow-up—but Kaplan-Meier analysis is not presented. The dePerrot study presents results after an extended transcervical approach with the interesting addition of video technology which they believe improves the complete resection rate. With this approach, they report a 41% CR rate with 4.1-year mean follow-up but only a 25% 5-year Kaplan-Meier estimated CR rate. These reports do not change the conclusions of our comparison between our results with extended TCT and those obtained by other groups with other approaches. With the reported crude cumulative CR rates for the different approaches overlapping, and with the only published Kaplan-Meier estimates (Jaretzki and Wolff [5], Shrager and associates [7], and current study) within a few percentage points of each other, we believe it is fair to say that the results do not dramatically differ according to approach.

We believe that CR rates with extended TCT have now been demonstrated to be sufficiently similar to those after the extended transsternal and "maximal" thymectomy procedures to allow one to argue cogently that whatever small difference may exist in CR rates is outweighed by the larger difference in morbidity between the different approaches. Further, the fact that most studies suggest that patients with less severe disease (our data included) and shorter preoperative duration of disease (not our data) enjoy greater CR rates lends some importance to providing referring neurologists with a less morbid operative approach they can feel comfortable recommending early to their patients.

It is important to note that there are two issues with our data that support the contention that "maximal" thymectomy does lead to slightly improved long-term outcomes. First, as previously mentioned, most studies, including ours, show a significant inverse association between severity of disease and response rates. The preoperative Osserman class in our series is somewhat lower (at 2.3) than in most other series of thymectomy, likely because the low morbidity associated with TCT leads neurologists to refer their patients with less severe disease—even isolated ocular disease. It is likely that this in part accounts for our CR rates being as high as they are.

Secondly, although our reanalysis of our earliest 84 patients shows that CR rates do not fall off with longer follow-up, it does appear to be the case that they do not continue to rise at the rate that Jaretzki and colleagues [5] report for the "maximal" thymectomy procedure. The latest data from the Jaretzki group show a Kaplan-Meier curve that rises from a CR rate of approximately 50% at 5 years to 81% at 7.5 years. Our curve, on the other hand, seems to become virtually flat after 6 years. It is of some concern to us that Jaretzki's "maximal thymectomy" curve does not appear to have been updated since a 1988 publication [5]. Updating with longer follow-up would allow determination of whether the Kaplan-Meier estimates have been confirmed and the curve does in fact continue to rise steeply. If so, it is conceivable that Jaretzki is correct in his contention that a more complete removal of thymic tissue by "maximal thymectomy" confers this advantage. In that case, this would need to be weighed against the increased morbidity of the operation.

Some have argued that the failure of a Kaplan-Meier CR curve after thymectomy to continue to rise in the later years of follow-up suggests that residual, immunologically active thymic tissue has been left in place [18]. We would suggest, however, that if residual tissue was a major issue, that the curve would not only fail to rise, but that it would probably actually fall as patients have disease relapses. That we do not see such a fall-off in our Kaplan-Meier curve (even when we reanalyze our earlier group of 84 patients operated upon before September 1999 with prolonged follow-up) suggests to us that, with extended TCT, we do not have a significant problem with functional, unresected thymic tissue.

Finally, it is worth noting that whereas in our first 78 patients we saw a significantly improved rate of CR in patients with small thymomas, analysis of this larger series shows only a trend in this direction. Although results in patients with thymoma are not statistically better than for nonthymomatous MG, the presence of a small, completely resected thymoma in stage 1 or 2 is clearly no reason to be pessimistic about the patient's possibility of recovery.

As Jaretzki and colleagues [18] have written, the optimal surgical technique of thymectomy in MG will be the one that "...balances extent of resection, morbidity, patient acceptance, and results."

Response rates after extended TCT are sufficiently high that, given its relative lack of morbidity, it is in our hands the procedure of choice in MG. Unfortunately, it is highly unlikely—owing to surgeons' real or claimed lack of equipoise on the issue—that a randomized trial of different techniques of thymectomy will ever be performed. We support the suggestion of the Clinical Research Standard Committee of the Myasthenia Gravis Foundation of America that multi-institutional, nonrandomized, prospective studies of two or more thymectomy techniques be organized as the best possible realistic alternative to a randomized study [19].

	Acknowledgments

Top Abstract Introduction Patients and Methods Results Comment Acknowledgments References

 
The authors would like to acknowledge the assistance of Julien Bonifacio in database management and Dr Jesus Gonzelez-Abraham who participated in data collection.

	References

Top Abstract Introduction Patients and Methods Results Comment Acknowledgments References

 

1. Blosson GB, Ernstoff RM, Howells GA, Bendick PJ, Glover JL. Thymectomy for myasthenia gravis Arch Surg 1993;128:855-862.[Abstract/Free Full Text]
2. Busch C, Machens A, Pichlmeier U, Emskotter T, Izbicki JR. Long-term outcome and quality of life after thymectomy for myathenia gravis Ann Surg 1996;224:225-232.[Medline]
3. Stern LE, Nussbaum MS, Quinlan JG, Fischer JE. Long-term evaluation of extended thymectomy with anterior mediastinal dissection for myasthenia gravis Surgery 2001;130:774-780.[Medline]
4. Schumacher ED, Roth J. Thymectomie bei einem fall von morbus basedowi mit myasthenie Mitteilg Grenzgebiete Med Chir 1912;25:746.
5. Jaretzki III A, Wolff M. "Maximal" thymectomy for myasthenia gravis: surgical anatomy and operative technique J Thorac Cardiovasc Surg 1988;96:711-716.[Abstract]
6. Calhoun RF, Ritter JH, Guthrie TJ, et al. Results of transcervical thymectomy for myasthenia gravis in 100 consecutive patients Ann Surg 1999;230:555-561.[Medline]
7. Shrager JB, Deeb ME, Mick R, et al. Transcervical thymectomy for myasthenia gravis achieves results comparable to thymectomy by sternotomy Ann Thorac Surg 2002;74:320-327.[Abstract/Free Full Text]
8. Nussbaum MS, Rosenthal GJ, Samaha FJ, et al. Management of myasthenia gravis by extended thymectomy with anterior mediastinal dissection Surgery 1992;112:681-688.[Medline]
9. Kas J, Kiss D, Simon V, et al. Decade-long experience with surgical therapy of myasthenia gravis: early complications of 324 transsternal thymectomies Ann Thorac Surg 2001;72:1691-1697.[Abstract/Free Full Text]
10. Jaretski III A, Barohn RJ, Ernstoff RM, et al. Myasthenia gravis: recommendations for clinical research standards Ann Thorac Surg 2000:327-334.
11. Cooper JD, Al-Jilaihawa AN, Pearson FG, Humphrey JG, Humphrey HE. An improved technique to facilitate transcervical thymectomy for myasthenia gravis Ann Thorac Surg 1988;45:242-247.[Abstract]
12. Miano MA, Bosley TM, Heiman-Patterson TD, et al. Factors influencing outome of prednisone dose reduction in myasthenia gravis Neurology 1991;41:919-921.[Abstract/Free Full Text]
13. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations J Am Stat Assoc 1958;53:457-481.
14. Mantel N. Evaluation of survival data and two new rank order statistics arising in its consideration Cancer Chemother Rep 1966;50:163-170.[Medline]
15. Cox DR. Regression models and lifetables J R Stat Soc B 1972;34:187-220.
16. Huang CS, Hsu HS, Huang BS, et al. Factors influencing the outcome of transsternal thymectomy for myasthenia gravis Acta Neurol Scand 2005;112:108-114.[Medline]
17. dePerrot M, Bril V, McRae K, Keshavjee S. Impact of minimally invasive transcervical thymectomy on outcome in patients with myasthenia gravis Eur J Cardiothorac Surg 2003;24:677-683.[Abstract/Free Full Text]
18. Jaretzki A, Steinglass KM, Sonnett JR. Thymectomy in the management of myasthenia gravis Semin Neurol 2004;24:49-62.[Medline]
19. Jaretzki A, Aarli JA, Kaminski HJ, Phillips LH, Sanders DB, Clinical Research Standards Committee, Medical/Scientific Advisory Board, Myasthenia Gravis Foundation of America Thymectomy for myasthenia gravis: evaluation requires controlled prospective studies Ann Thorac Surg 2003;76:1-3.[Free Full Text]

This article has been cited by other articles:

				C. Cheng, G. Wu, S.-C. J. Yeung, R. Li, A. E. Bella, J. Pang, F.-t. Zhong, H. Luo, Y. Jin, and J. Pan Serum protein profiles in myasthenia gravis. Ann. Thorac. Surg., October 1, 2009; 88(4): 1118 - 1123. [Abstract] [Full Text] [PDF]

				E. Meacci, A. Cesario, S. Margaritora, V. Porziella, A. Tessitore, G. Cusumano, A. Evoli, and P. Granone Thymectomy in myasthenia gravis via original video-assisted infra-mammary cosmetic incision and median sternotomy: long-term results in 180 patients Eur. J. Cardiothorac. Surg., June 1, 2009; 35(6): 1063 - 1069. [Abstract] [Full Text] [PDF]

				M. Lucchi, R. Ricciardi, F. Melfi, L. Duranti, F. Basolo, G. Palmiero, L. Murri, and A. Mussi Association of thymoma and myasthenia gravis: oncological and neurological results of the surgical treatment Eur. J. Cardiothorac. Surg., May 1, 2009; 35(5): 812 - 816. [Abstract] [Full Text] [PDF]

				P. Gadient, J. Bolton, and V. Puri Juvenile Myasthenia Gravis: Three Case Reports and a Literature Review J Child Neurol, May 1, 2009; 24(5): 584 - 590. [Abstract] [PDF]

				A. Chaudhuri and P.O. Behan Myasthenic crisis QJM, February 1, 2009; 102(2): 97 - 107. [Abstract] [Full Text] [PDF]

				M. Zielinski Patients With Myasthenia With Oropharyngeal Involvement Need Medical Preparation Before Thymectomy Ann. Thorac. Surg., May 1, 2008; 85(5): 1842 - 1842. [Full Text] [PDF]

				A. Jaretzki III and J. R. Sonett Evaluation of Results of Thymectomy for MG Requires Accepted Standards Ann. Thorac. Surg., July 1, 2007; 84(1): 360 - 361. [Full Text] [PDF]

				J. B. Shrager and L. R. Kaiser Reply. Ann. Thorac. Surg., July 1, 2007; 84(1): 361 - 361. [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Personal Folders Download to citation manager Author home page(s): Joseph B. Shrager Zhen Chen Larry R. Kaiser Permission Requests Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Shrager, J. B. Articles by Kaiser, L. R. Search for Related Content PubMed PubMed Citation Articles by Shrager, J. B. Articles by Kaiser, L. R. Related Collections Mediastinum