Annals of Surgical Oncology Cite Track
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
10.1245/s10434-006-9233-3
Annals of Surgical Oncology 14:2113-2120 (2007)
© 2007 Society of Surgical Oncology
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Kimmel, Z.
Right arrow Articles by Alam, M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Kimmel, Z.
Right arrow Articles by Alam, M.

Original Article

Peripheral Excision Margins for Dermatofibrosarcoma Protuberans: A Meta-analysis of Spatial Data

Zebadiah Kimmel, MS1, Desiree Ratner, MD2, John Y. S. Kim, MD3, Jeffrey D. Wayne, MD4, Alfred W. Rademaker, MD, PhD5 and Murad Alam, MD6,7

1 Feinberg School of Medicine, Northwestern University, Chicago, IL, USA
2 Department of Dermatology, College of Physicians and Surgeons, Columbia University, New York, NY, USA
3 Division of Plastic Surgery, Northwestern University, Chicago, IL, USA
4 Division of Surgical Oncology, Northwestern University, Chicago, IL, USA
5 Department of Preventive Medicine (Biostatistics), Northwestern University, Chicago, IL, USA
6 Department of Dermatology, Northwestern University, 675 North St. Clair Street, Suite 19-150, Chicago, IL 60611, USA
7 Department of Otolaryngology-Head and Neck Surgery, Northwestern University, Chicago, IL, USA

Correspondence: Address correspondence and reprint requests to: Murad Alam, MD; E-mail: m-alam{at}northwestern.edu


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 CONCLUSIONS
 REFERENCES
 
Background: The purpose of this study was to analyze available datasets to assess the degree of asymmetry typically associated with DFSP, and to study the optimal surgical approach for extirpating these tumors by clearing lateral margins.

Methods: MEDLINE (1994–2004) was searched for English-language multi-patient series concerning DFSP. Case series were included if complete information could be obtained for: (a) two-dimensional preoperative tumor size as measured on the skin surface before removal; (b) postoperative tumor size, as estimated by the dimensions of the final wound defect. Four case series met these criteria, and the authors were contacted directly for unpublished raw data.

Results: For each of 98 included tumors we computed: (1) the tumor index, a measure of clinically apparent tumor surface area; (2) clearance margin, or the theoretical best and worst-case surgical margins that would have been required for tumor clearance. We used this information to (a) assess the relationship between clinically apparent tumor size and final surgical margin; (b) determine the proportion of tumors of a given size that would be cleared by a margin of given width. We found that standard wide excision margin of 4 cm was predicted to provide a tumor clearance rate of 95% for tumors of preoperative size less than or equal to 3 cm x 3 cm.

Conclusions: There is a weak relationship between preoperative tumor size and the width of the final defect after clearance. Based on our calculations, very wide local excision is necessary for clearance of most DFSPs.

Key Words: Dermatofibrosarcoma protuberans • Excision • Margins • Mohs • Interdisciplinary


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 CONCLUSIONS
 REFERENCES
 
Dermatofibrosarcoma protuberans (DFSP) is an uncommon, slow-growing dermal tumor that is locally invasive and, rarely, metastatic.1 It has an unpredictable growth pattern, characterized by deeply and extensively infiltrating tendrils of tumor permeating laterally into the far subcutis.2 Even a DFSP that appears relatively small may in fact have contiguous thin, distant projections that intercalate within fat lobules.

In recent years, lesions of DFSP have been excised by Mohs micrographic surgery (MMS),3 or by local wide excision, usually with a margin of 3 cm,4 although margins of up to 5 cm have been suggested.5

The primary question that this study addresses is whether, for a DFSP tumor of any given preoperative size, there is an appropriate excisional margin that should be used to clear the tumor. The secondary question that this study addresses is whether a distinction can be made between larger and smaller tumors for the purposes of surgical management.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 CONCLUSIONS
 REFERENCES
 
Clinical Material
We conducted a MEDLINE search for multi-patient series within the past ten years in which DFSP was excised with Mohs micrographic surgery or similar three-dimensional frozen section technique that visualized the entire peripheral and deep margin of the surgical specimen during the process of tumor removal. Because such a surgical approach entails removal of successive stages until clearance is obtained, final wound defect data is a relatively accurate estimate of histological tumor size. We obtained raw data, consisting of preoperative and postoperative lesion size; if this information did not explicitly appear in a study, we attempted to contact the authors to retrieve it. We were able to obtain raw data from four studies,6–9 for a total of 98 data points (Table 1Go). In one additional study we obtained raw data but it could not be used, as tumor size was provided in one dimension only.10 We were unable to obtain raw data from four studies.2,3,11–12


View this table:
[in this window]
[in a new window]

  		TABLE 1. Preoperative tumor extent and postoperative tumor size of primary and recurrent DFSPs (X and Y represent width and length, in cm; tumor index is computed from preoperative measurements)
 
All the tumors we studied were excised with intraoperative frozen section analysis of lateral and deep margins. If initial margins were positive by frozen section, more tissue was harvested along the positive edge, and further frozen section analysis was performed until all peripheral and deep margins were negative. Hence the assumption was that there was no discrepancy between the final frozen section margins and the final surgical margins. In 3 of 98 cases, tumors were reported to recur within the follow-up periods of their respective studies (Table 1Go), and re-excision was required. For each of these instances, we assumed that the final frozen section margins had missed tumor at the margin, and the true histological extent of the tumors was greater than the post-surgical extent after the first surgery. Since final surgical margins after the re-excisions were not available, we continued to use final surgical margins after the initial surgeries as a proxy under the assumptions that: (a) these were close in absolute extent to final surgical margins; (b) that given that final surgical margins were at least slightly greater than initial surgical margins, using the latter in modeling would be the most statistically conservative choice, leading to the lowest estimates of margins required for tumor clearance.

Calculation of Surgical Margins
For each DFSP in this analysis, we estimated the surgical margin that would have been required for clearance if the tumor had instead been treated with wide excision.8 For each tumor, a preoperative lesion size and a postoperative defect size was obtained. The preoperative lesion was defined as the extent of the clinically apparent tumor, before any treatment was undertaken; the preoperative lesion definition specifically excluded any clinically inapparent tumor mass. The postoperative defect was defined as the surgical defect that remained once the surgical removal was completed; if the surgery had been successful, then the postoperative defect by definition encompassed the entire extent of the tumor. The tissue lying between the boundaries of the preoperative and postoperative defects was defined as the clinically inapparent tumor mass. The basis for our analysis was the observation that the maximum (radial) difference between the boundaries of the preoperative and postoperative defects would be equal to the minimum surgical margin that would have been required to clear the tumor, had wide excision been performed instead of excision with intraoperative frozen section analysis of lateral margins. This is shown schematically in Fig. 1Go.


Figure 1
View larger version (13K):
[in this window]
[in a new window]

  		FIG. 1. Schematic representation of preoperative lesion size and postoperative defect size under the assumptions of concentric and eccentric tumor growth.

 
If a DFSP grows purely concentrically, then the preoperative lesion would lie exactly at the center of the postoperative defect. If a DFSP grows purely eccentrically, then the preoperative lesion would abut the edge of the postoperative defect. In either case, the surgical margin that would be required for wide excision would be equal to the maximum (radial) distance between the two lesion boundaries. It is clear from Fig. 1Go that this margin would be narrowest when growth is purely concentric, and widest when growth is purely eccentric. We will refer to the surgical margin under the assumption of purely concentric growth as the best-case margin, and to the surgical margin under the assumption of purely eccentric growth as the worst-case margin. It should be noted that the best-case margin would be equal to one-half of the worst-case margin (see Fig. 1Go).

We defined E, the "eccentricity coefficient", as a quantity between 0 and 1 for any particular tumor that denoted the geometric growth pattern of this tumor. An eccentricity coefficient of 0 would indicate purely concentric growth; an eccentricity coefficient of 1 would indicate purely eccentric growth. E may be measured for any tumor by measuring the distance between the centroids of the preoperative lesion and the postoperative defect. (The centroid of a shape is its center of gravity. We define the eccentricity coefficient {varepsilon} = the measured distance between the centroids of the preoperative and postoperative defects/ the theoretical distance between the centroids of the preoperative and postoperative defects had the tumor grown purely eccentrically.) Such a measurement is not routinely performed during surgery.

The true margin of a given tumor was defined as the minimum surgical margin that would have been required to completely excise it, had its true extent been evident preoperatively. If E were equal to zero, then the true margin would be the same as the best-case margin. If E were equal to one, then the true margin would be the same as the worst-case margin.

Based on these assumptions and nomenclature, if X and Y represent the measured length and width of a lesion, then:


Formula


Formula


Formula

Calculation of Tumor Index
We found it useful to categorize initially apparent tumor size using the derivative construct of "tumor index," which varied with surface area of the tumor. Specifically, the tumor index, T, was defined as the square root of (X2 + Y2). The utility of this approach is intuitively obvious since a tumor of size 3 cm x 4 cm requires surgical management comparable to that of a tumor of size 4 cm x 3 cm. To extend this example, the tumor index of a 3 cm x 4 cm (or 4 cm x 3 cm) tumor would be 5. Tumors of equal surface area may therefore be grouped into sets with similar tumor indices.

Biostatistical Supervision
Data extraction and data analysis were supervised by Dr. Rademaker, director of the biostatistics core at the Robert E. Lurie Comprehensive Cancer Center at Northwestern University. Dr. Rademaker reviewed the mathematical model created to estimate theoretical margins for tumor clearance. He and his staff also performed statistical tests (Pearson correlation and Fisher exact test) to assess this model.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
CONCLUSIONS
 REFERENCES
 
Tumor Size versus Margins
Figure 2Go displays all best-case margins for removal of tumors of various widths and lengths in the form of a continuous landscape. Figure 3Go shows tumor index versus best-case and worst-case surgical margins, with a green circle representing the best-case margin, and a red cross, worst-case margin. For tumors no larger than 3 cm x 3 cm in length and width (tumor index of approximately 4), wide local excision of 4 cm around the clinically apparent margins would have been sufficient to histologically clear 95% of such tumors.


Figure 2
View larger version (43K):
[in this window]
[in a new window]

  		FIG. 2. Landscape plot of best-case margins (cm) required for tumor clearance as a function of preoperative tumor size.

 

Figure 3
View larger version (17K):
[in this window]
[in a new window]

  		FIG. 3. Best-case (green circles) and worst-case (red crosses) margins (cm) required for tumor clearance as a function of tumor index.

 
Cursory examination of these figures suggests there is no strong relationship between clinically apparent tumor size (either XY measurements, or tumor index) and the surgical margin required for wide excision. Tumors that appear small clinically may require large margins for excision, and vice versa. This is best appreciated by noticing that the landscape of Fig. 2Go is hilly throughout its extent. The landscape does not slope upward with increasing tumor size, as one would intuitively expect.

Correlation of Apparent Clinical Extent of Tumor to Excision Margin
A statistical analysis was performed to assess the correlation between preoperative tumor size and width of the margin required for clearance. For the best-case scenario, the Pearson correlation coefficient was 0.60, and for the worst-case scenario, the correlation coefficient was 0.55. While both of these are statistically significant (p < 0.01), the strengths of both correlations are well below the threshold of 0.8 or 0.9 regarded by many as a benchmark for a high level of correlation in biologic systems.

Comparison of Excision Margin for Tumors Grouped by Initial Tumor Index
Further statistical analysis entailed the use of Fisher’s exact test to measure the association between initial tumor index and margin required for clearance. Statistics were computed for both best case (concentric growth) and worst-case (eccentric growth) scenarios. Given the limited numbers of tumors, the lesions were grouped into categories: tumor index 1–2; tumor index 3–5; tumor index 6 or greater. Margins were assessed to within 0.5 cm, and margin categories were 1–2 cm; 2.5–4 cm; 4.5 cm or greater.

In the best-case scenario, the test statistic (p = 0.0472) indicated that there was a statistically significant relationship between preoperative tumor index and the margin required for clearance. Overall, larger tumors tended to require greater margins for clearance. In the worst-case scenario, however, the statistic (p = 0.1362) indicated no significant relationship between tumor size and margin. Even in the best-case scenario, the difference was only statistically significant when the tumors and margins were grouped into a few major categories, confirming a weak relationship overall between clinically apparent tumor size and true histologic extent.

Empirical Estimation of Eccentricity of Tumors
Given that we computed best- and worst-case scenarios based on possible variations in the underlying eccentricity of tumor growth, it was important to determine which assumption was a better approximation of real tumor growth. We attempted to clarify whether DFSPs grow concentrically or eccentrically by estimating the average value of E for DFSPs. We estimated the real-world value of the eccentricity coefficient E by comparing predicted recurrence rates for different values of E (E = 0 or 1) against real-world recurrence rates. We found relatively close agreement between best-case predictions and real-world rates, suggesting that E is a small quantity (Table 2Go).


View this table:
[in this window]
[in a new window]

  		TABLE 2. Comparison of best-case and worst-case recurrence rates predicted for DFSPs (n = 98) treated with wide local excision
 

   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 CONCLUSIONS
REFERENCES
 
Relationship between Clinically Apparent Size and True Extent
We reviewed 98 DFSPs, stratified by tumor index, and calculated the corresponding margins that would have been required for clearance had wide excision been employed instead of intraoperative frozen section-guided excision. We found at best a modest correlation between clinically apparent tumor size and requisite surgical margin for DFSPs. This suggested that preoperative tumor size did not correlate well with true tumor extent. Even when tumors and margins were grouped into a few major categories, a relationship between apparent size and surgically verified margin was only seen in the best-case scenario of perfectly concentric tumor growth around the clinically apparent lesion.

Standard Width of Surgical Margin Required
A surgical margin of 3 cm is currently accepted as the standard of care for excision of DFSP. Based on our analysis, this margin actually provides inadequate tumor clearance, as fewer than 90% of presenting lesions would have been completely removed. Because of the high risk of recurrence of DFSP, a 3 cm margin should probably be used with great caution and with appropriate follow-up for early detection of any recurrence. While our results suggest that a margin of 4 cm is expected to clear 95% of DFSP of small to moderate size, it is important to note that a margin of this size is likely to result in unnecessary removal of large amounts of normal tissue.

Recommendations for Surgical Management of DFSP
Our analysis indicates that clinically apparent size of DFSP is a poor indicator of its true histologic extent, and that wider than expected margins are often required to clear DFSPs of all sizes. Taken together, these results strongly suggest that careful histological analysis of tissue specimens is essential to ensure high surgical cure rates. Wide local excision with a 4 cm margin may be impractical at many anatomic sites. For smaller apparent tumors of DFSP, intraoperative frozen-section analysis of peripheral/lateral margins in the subcutaneous space can thus reduce the likelihood of recurrence without creating a needlessly massive defect. For very large tumors, peripheral cutaneous margins may be obtained by Mohs,13 and deep margins by surgical oncology for permanent section analysis. This combination multidisciplinary approach may include staged frozen-section analysis of skin and subcutis around the central tumor, followed by extirpation and repair of the central deeply penetrating nodule by the joint efforts of multiple surgical services, including surgical oncology and plastic surgery. Preoperative use of imatinib may be considered as an adjuvant treatment to decrease the clinical size of massive tumors prior to definitive surgical excision.14 Thus, rather than the traditional approach of obtaining uniform 3 cm margins around apparent lesions of DFSP, either Mohs micrographic surgery or a combination multidisciplinary approach15,16 that includes detailed analysis of histopathologic margins may be the most appropriate modalities for the treatment of DFSP of any size (Fig. 4Go). At Northwestern, large DFSPs are treated as follows: peripheral margins to deep fat are cleared by Mohs surgery that entails removal of a ring of tissue around the central tumor core; subsequently, surgical oncology clears deep margins, typically to below fascia in superficial muscle; and once frozen-section analysis verifies a tumor-free deep plane, flap or graft repair is performed by plastic surgery (Fig. 5Go). Notably, DFSP seldom penetrates the muscular fascia and usually uncertainty regarding excision margin pertains to the peripheral margins.


Figure 4
View larger version (32K):
[in this window]
[in a new window]

  		FIG. 4. Tumor size, extent, and anatomic location may be considered when assembling a surgical team to address a given DFSP.

 

Figure 5
View larger version (37K):
[in this window]
[in a new window]

  		FIG. 5. At least one week may be required for a typical team comprised of a dermatologist, a surgical oncologist, and a plastic surgeon may remove a large DFSP definitively.

 
Whether frozen-section margins are analyzed by dermatologists or surgical pathologists can be a contentious issue, but the challenging histological features of DFSP encourage cooperation rather than competition. At Northwestern, harvesting and frozen-section analysis of peripheral margins is performed by a dermatologist with subspecialty training in cutaneous oncology until clear peripheral margins are obtained. In our experience, in at least 10% of cases, artifact on frozen-section analysis of high-risk peripheral margins results in tumor being missed, so once the surgical dermatologist has obtained supposedly tumor-free peripheral margins, an additional 1–2 mm circumferential peripheral margin is obtained, mapped, divided, inked en-face, and sent for rush permanent section processing to the dermatopathology laboratory. This peripheral margin is then read by a dermatopathologist, and if it is positive, further frozen sections are obtained by the surgical dermatologist at the site of positivity. Once peripheral margins have been verified as clear on permanent sections by the dermatopathologist, the deep margins are harvested by surgical oncology. The base of the deep margin is inked and sent to surgical pathology. While the first deep margin obtained immediately below fascia is usually devoid of tumor, if tumor is found, the surgical oncologist obtains a further margin and sends this to surgical oncology for similar processing. This time-intensive process of three different services (surgical dermatology, dermatopathology, and surgical pathology) checking DFSP margins must be explained to the patient prior to surgery. Patients need to understand that the tumor’s filamentous projections, particularly in the peripheral plane, require a methodical, step-wise, and redundant approach to mimimize the risk of recurrence. While DFSP rarely metastasizes, recurrences are common, and given the frequently large size of these tumors, should be aggressively avoided to reduce the likelihood of morbidity and functional loss. In conclusion, excision of any DFSP regardless of the surgical modality used17 may result in defects of considerable size due to the typically wide and deep subclinical histological extension of this tumor. Multidisciplinary management is most often indicated. Specifically, involvement of experts in general surgery, plastic surgery, dermatology, otolaryngology, orthopedics, and/or medical oncology, may be appropriate in the context of extensive disease.

Received for publication May 28, 2006. Accepted for publication June 5, 2006.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 CONCLUSIONS
 REFERENCES
 

  1. Laskin WB. Dermatofibrosarcoma protuberans. CA Cancer J Clin 1992; 42:116–25.[Abstract]
  2. Haycox CL, Odland PB, Olbricht SM, Casey B. Dermatofibrosarcoma protuberans (DFSP): growth characteristics based on tumor modeling and a review of cases treated with Mohs micrographic surgery. Ann Plast Surg 1997; 38:246–51.[Medline]
  3. Gloster HM Jr, Harris KR, Roenigk RK. A comparison between Mohs micrographic surgery and wide surgical excision for the treatment of dermatofibrosarcoma protuberans. J Am Acad Dermatol 1996; 35:82–7.[CrossRef][Medline]
  4. Kricorian GJ, Schanbacher CF, Kelly AP, Bennett RG. Dermatofibrosarcoma protuberans growing around plantar apo-neurosis: excision by Mohs micrographic surgery. Dermatol Surg 2000; 26:941–5.[CrossRef][Medline]
  5. Arnaud EJ, Perrault M, Revol M, et al. Surgical treatment of dermatofibrosarcoma protuberans. Plast Reconstr Surg 1997; 100:884–95.[CrossRef][Medline]
  6. Dawes KW, Hanke CW. Dermatofibrosarcoma protuberans treated with Mohs micrographic surgery: cure rates and surgical margins. Dermatol Surg 1996; 22(6):530–4.[CrossRef][Medline]
  7. Hafner J, Schutz K, Morgenthaler W, et al. Micrographic surgery (,slow Mohs’) in cutaneous sarcomas. Dermatology 1999; 198:37–43.[CrossRef][Medline]
  8. Ratner D, Thomas CO, Johnson TM, et al. Mohs micrographic surgery for the treatment of dermatofibrosarcoma protuberans. Results of a multiinstitutional series with an analysis of the extent of microscopic spread. J Am Acad Dermatol 1997; 37:600–13.[CrossRef][Medline]
  9. Tom WD, Hybarger CP, Rasgon BM. Dermatofibrosarcoma protuberans of the head and neck: treatment with Mohs surgery using inverted horizontal paraffin sections. Laryngoscope 2003; 113:1289–93.[CrossRef][Medline]
  10. Parker TL, Zitelli JA. Surgical margins for excision of dermatofibrosarcoma protuberans. J Am Acad Dermatol 1995; 32:233–6.[CrossRef][Medline]
  11. Ah-Weng A. Dermatofibrosarcoma protuberans treated by micrographic surgery. Br J Cancer 2002; 87:1386–9.[CrossRef][Medline]
  12. Nouri K, Lodha R, Jimenez G, Robins P. Mohs micrographic surgery for dermatofibrosarcoma protuberans: University of Miami and NYU experience. Dermatol Surg 2002; 28:1060–4.[CrossRef][Medline]
  13. Moossavi M, Alam M, Ratner D. Use of the double-bladed scalpel in peripheral margin control of dermatofibrosarcoma protuberans. Dermatol Surg 2000; 26:599–601.[CrossRef][Medline]
  14. McArthur GA, Demetri GD, van Oosterom A, Heinrich MC, Debiec-Rychter M, Corless CL, et al. Molecular and clinical analysis of locally advanced dermatofibrosarcoma protuberans treated with imatinib: Imatinib Target Exploration Consortium Study B225. J Clin Oncol 2005; 23:866–73.[Abstract/Free Full Text]
  15. DuBay D, Cimmino V, Lowe L, Johnson TM, Sondak VK. Low recurrence rate after surgery for dermatofibrosarcoma protuberans: a multidisciplinary approach from a single institution. Cancer 2004; 100:1008–16.[CrossRef][Medline]
  16. Wacker J, Khan-Durani B, Hartschuh W. Modified Mohs micrographic surgery in the therapy of dermatofibrosarcoma protuberans: analysis of 22 patients. Ann Surg Oncol 2004; 11:438–44.[Abstract/Free Full Text]
  17. Snow SN, Gordon EM, Larson PO, Bagheri MM, Bentz ML, Sable DB. Dermatofibrosarcoma protuberans: a report on 29 patients treated by Mohs micrographic surgery with long-term follow-up and review of the literature. Cancer 2004; 101:28–38.[CrossRef][Medline]



This article has been cited by other articles:


Home page
 Ann. Surg. Oncol.Home page
J. S. Zager, M. M. Bui, J. M. Farma, J. L. Messina, and V. K. Sondak
Proper Margins of Excision in Dermatofibrosarcoma Protuberans: Wide or Narrow?
Ann. Surg. Oncol., September 1, 2008; 15(9): 2614 - 2616.
[Full Text] [PDF]

Home page
 Ann. Surg. Oncol.Home page
Z. Kimmel and M. Alam
Peripheral Excision Margins for Dermatofibrosarcoma Protuberans
Ann. Surg. Oncol., September 1, 2008; 15(9): 2617 - 2617.
[Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Kimmel, Z.
Right arrow Articles by Alam, M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Kimmel, Z.
Right arrow Articles by Alam, M.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS