This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Chu, Q. D. Articles by Gibbs, J. F. Search for Related Content PubMed PubMed Citation Articles by Chu, Q. D. Articles by Gibbs, J. F. Related Collections Prognostic factors

Predictive Factors Associated With Long-Term Survival in Patients With Neuroendocrine Tumors of the Pancreas

From the Department of Surgical Oncology, Roswell Park Cancer Institute, State University of New York at Buffalo, Buffalo, New York.

Correspondence: Address correspondence and reprint requests to: John F. Gibbs, MD, Department of Surgical Oncology, Roswell Park Cancer Institute, Elm and Carlton Streets, Buffalo, NY 14263; Fax: 716-845-2320; E-mail: john.gibbs{at}roswellpark.org

	ABSTRACT

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
Background: Neuroendocrine tumors of the pancreas are rare tumors. We identified predictive factors that are associated with long-term survival (5 years).

Methods: Fifty patients with a diagnosis of neuroendocrine tumors of the pancreas were retrospectively evaluated. The following factors were evaluated for disease-specific mortality: age, sex, primary tumor location, functional status, type of primary tumor treatment, presence or absence of liver metastases, timing of liver metastases occurrence, and type of liver metastases treatment. Aggressive treatment of the liver metastases included surgery, chemoembolization, or intrahepatic arterial infusion chemotherapy.

Results: Twenty-three patients (47%) had tumor located in the head of the pancreas, and 29 patients (58%) had nonfunctioning tumor. Thirty-nine patients (78%) had liver metastases. The median follow-up for the entire group was 35 months (range, .76–206 months). The median survival for the entire group was 40 months, and the overall 1-, 2-, and 5-year survival rates were 84%, 69%, and 36%, respectively. Factors that had a significant favorable effect on survival included curative resection of the primary tumor, metachronous liver metastases, absence of liver metastases, and aggressive treatment of the liver metastases.

Conclusions: Definitive surgical resection of the primary tumor, absence of liver metastases, metachronous liver metastases, and aggressive treatment of the liver metastases were predictors of long-term survival in patients with neuroendocrine tumors of the pancreas.

Key Words: Neuroendocrine tumors • Pancreas • Liver metastases • Predictive factors

	INTRODUCTION

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
Neuroendocrine tumors involving the pancreas are a rare entity affecting less than one person per 100,000 population.¹ One case is seen for every 125 patients with adenocarcinoma of the pancreas.² However, unlike adenocarcinoma of the pancreas, neuroendocrine tumors of the pancreas carry a better prognosis. The natural history of this uncommon entity has not been clearly defined. Many patients present with synchronous metastases, typically to the liver. A common dilemma arises concerning the treatment of these patients. Should the treating physician embark on an aggressive approach, which may include pancreatectomy and metastasectomy, or treat these patients expectantly, with a conservative approach? The purpose of this study was to identify predictive factors that are associated with long-term survival (5 years) in patients with neuroendocrine tumors of the pancreas. Specifically, we evaluated the role of aggressive intervention in the context of multidisciplinary patient management.

	PATIENTS AND METHODS

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
The medical records of 50 patients admitted to Roswell Park Cancer Institute from January 1, 1970, to June 30, 2001, with a diagnosis of neuroendocrine tumor of the pancreas were retrospectively reviewed. The Tumor Registry data, clinic notes, and medical records for these patients were reviewed. All patients underwent a complete history and physical examination and laboratory and radiological investigations, and all had cytological or pathologic confirmation of neuroendocrine tumors.

Abdominal computed tomography scans were obtained in all but four patients; these four patients had a liver/spleen scintigraphy as part of their radiological work-ups. Surveillance abdominal computed tomography scans were obtained every 6 months to identify progression of disease.

Tumors were classified as functioning neuroendocrine tumors on the basis of clinical symptoms of hormonal excess accompanied by increased serum peptide levels. Tumors were classified as nonfunctioning if no clinical symptom existed, even in the presence of increased serum peptide or immunohistochemistry positivity of gastrointestinal hormones. Metachronous liver metastases were defined as lesions that were identified by radiological imaging 1 year after the date of the initial surgery.

Aggressive treatment of the liver metastases included surgery, chemoembolization, or continuous intrahepatic arterial chemotherapy infusion (IHAC). Nonaggressive treatments included observation or systemic chemotherapy. Patients who were selected for IHAC or chemoembolization must have had an Eastern Cooperative Oncology Group performance status of 0 to 2, no evidence of extrahepatic disease, no evidence of systemic infection, and no evidence of portal vein thrombosis.

Descriptive statistics were calculated by using means, medians, and frequencies as appropriate to the type of data. All times were reported in months and calculated from the date of treatment to a specified end point (i.e. death, date of last follow-up). Disease-specific mortality was defined as tumor-related death. Patients who died of other causes, regardless of disease status, were censored at the time of death in the analysis of disease-specific mortality. The end point for the univariate analyses was disease-specific survival. Follow-up was obtained through medical records, telephone contacts, or letters. Patient follow-up was complete up to June 30, 2001. The following factors were evaluated by univariate analysis to determine their effect on long-term survival: sex, location of the primary tumor (head of the pancreas vs. body/tail of the pancreas), tumor functional status, type of treatment of the primary tumor, presence or absence of liver metastases, time of occurrence of the liver metastases, and the type of treatment of the liver metastases.

Survival curves were generated by using the Kaplan-Meier method, ³ and the log-rank test⁴ was used to compare survival curves. Significance was defined as P < .05. All figures refer to disease-specific survival.

	RESULTS

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
Patient Characteristics
There were 28 men (56%) and 22 women (44%). The median age was 57 years (range, 27–77 years). One patient had a history of multiple endocrine neoplasia syndrome. Table 1 depicts the clinical presentation of the 50 patients. Patients with functioning tumors presented with symptoms of hormonal excess associated with increases of the corresponding peptides. The most common symptoms for both functioning and nonfunctioning tumors were abdominal and/or back pain and weight loss. Nausea and vomiting were more common in patients with nonfunctioning tumors than in patients with functioning tumors. Dermatitis, symptoms of dizziness, and syncopal episodes were seen in the functioning group but not in the nonfunctioning group.

Distant Metastases
Liver metastases occurred in 39 (78%) of 50 patients. Twenty-nine (74%) of 39 patients presented with synchronous liver metastases, whereas 10 patients presented with metachronous liver metastases. The median time for the development of metachronous liver metastases was 51 months (range, 16–96 months). Twelve (57%) of 21 patients with functioning tumors and 17 (59%) of 29 patients with nonfunctioning tumors presented with synchronous liver metastases. Three (14%) of 21 patients with functioning tumors and 7 (24%) of 29 patients with nonfunctioning tumors presented with metachronous liver metastases. Overall, 15 (71%) of 21 patients with functioning tumors and 24 (83%) of 29 patients with nonfunctioning tumors had liver metastases (Table 3).

Treatment Outcome
Surgical Management of the Primary Tumor
Surgical resection varied widely depending on the extent of organ involvement. Thirty-one patients (62%) had surgical exploration. Of these, 21 patients (42%) had potentially curative resections, 3 patients (6%) had intestinal bypasses, and 7 patients (14%) had exploration with biopsy only (Table 4). Resection of the primary tumor with en-bloc adjacent organ resection was performed in 12 (57%) of the 21 patients. One patient presented with a clinical diagnosis of an incarcerated groin hernia but was found to have extensive lymphadenopathy secondary to cancer metastasis. This patient underwent a groin dissection followed by a Whipple procedure 6 weeks later for a 6-cm primary tumor. Two (10%) of the 21 patients had concomitant liver resections (central hepatectomy and left lateral segmentectomy). Of the six patients with insulinoma, only one had enucleation for a 1.5-cm lesion; two had distal pancreatectomy and splenectomy, and three patients did not have their primary lesion resected because they had synchronous diffuse liver metastases.

View larger version (9K): [in this window] [in a new window]   FIG. 1. Disease-specific survival for all patients.

View larger version (11K): [in this window] [in a new window]   FIG. 2. Disease-specific survival comparing patients who underwent definitive resection of the primary tumor and those who did not.

View larger version (12K): [in this window] [in a new window]   FIG. 3. Disease-specific survival comparing patients with liver metastases and those without liver metastases.

View larger version (12K): [in this window] [in a new window]   FIG. 4. Disease-specific survival comparing patients who had aggressive treatment versus conservative treatment of liver metastases.

View larger version (11K): [in this window] [in a new window]   FIG. 5. Disease-specific survival comparing synchronous (syn) and metachronous (met) liver metastases.

Comparison of survival data according to age, sex, location of the primary tumor, and tumor functional status found no differences in overall disease-specific survival. There were three in-hospital deaths (6%) due to advanced disease. The overall morbidity was 18% (n = 9 patients). Three patients had grade 2 mucositis; three patients had nausea, fatigue, and diarrhea; and one patient each had a right subclavian vein thrombosis from a central line placement, pleural effusion, and radiation duodenitis. The patients who had chemoembolization experienced mild to moderate abdominal pain that was relieved by a patient-controlled analgesic pump.

	DISCUSSION

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
The treatment of patients with neuroendocrine tumors of the pancreas requires a thorough knowledge of its natural history. A multidisciplinary approach should be considered for these relatively slow-growing tumors because patients may effectively be cured or palliated over a period of years by multimodality therapies.

Our experience demonstrates that an aggressive approach should be considered in selected patients who present with neuroendocrine tumors of the pancreas. Liver involvement will occur in most patients. When feasible, primary resection of the primary tumor or liver metastases may be beneficial. In those patients who present with diffuse liver metastases but without evidence of extrahepatic disease and with a good performance status, an aggressive approach such as intra-arterial chemotherapy or chemoembolization should still be considered. Patients who have no evidence of liver metastases or those with metachronous liver metastases have a longer survival than those with liver disease or with synchronous liver metastases. Functioning and nonfunctioning tumors have similar frequency of liver metastases and survival.

Neuroendocrine tumors of the pancreas are classified as either functioning or nonfunctioning islet cell tumors. Because nonfunctioning tumors are morphologically indistinguishable from their functional counterparts, this differentiation relies on the presence or absence of symptoms associated with excessive secretion of hormones. Recent evidence, based on electron microscopy studies and immunohistochemical staining, suggests that nonfunctioning tumors do indeed elaborate hormones.⁵ It has been speculated that the reasons why patients are asymptomatic may be insufficient amounts of peptide production, release of biologically inactive molecular forms of the peptides, or adequate amounts of peptides produced but inadequate amounts released.⁶ The use of neuron-specific enolase as a marker for nonfunctioning islet cell carcinoma has been suggested by Prinz and Marangos.⁷

Some authors believe that the dramatic symptoms associated with functioning tumors may lead to earlier patient presentation and therefore a survival advantage over patients with nonfunctioning tumors.⁸ Broughan et al.⁸ reviewed their experience with 84 cases of pancreatic islet cell tumors and demonstrated a significantly lower 5-year survival of 63% for nonfunctioning tumors versus 97% for insulinoma and 68% for gastrinoma. Other investigators have demonstrated that the overall prognosis for both is similar.^9,10 In a review of 64 cases, Lo et al.⁹ also found no difference in survival between functioning and nonfunctioning tumors. Our own series demonstrated that the survivals for functioning versus nonfunctioning tumors were similar, although the small number in this report may have affected this analysis.

One might expect that patients with functioning tumors would present with smaller tumors and therefore be less likely to present with symptoms of tumor encroachment. However, it is interesting to note that our experience demonstrated that the median tumor size on presentation was identical for both functioning and nonfunctioning tumors (6 cm). In addition, the most common presentations for both groups were abdominal and/or back pain and weight loss. Such similarities in presentation and outcome, whether or not the tumor was functional, suggest that the classification based on functionality may not be as relevant as once thought. Indeed, White et al.¹⁰ demonstrated no significant difference with respect to the incidence of metastatic disease on presentation, resectability rate with curative intent, or disease-free survival between functioning islet cell tumors and nonfunctioning islet cell tumors.

Surgery remains the cornerstone of treatment. In our series, patients who had a complete resection of the primary lesion had a significantly longer survival than those who did not. This was also observed by other investigators.^11,12–14 Evans et al.¹¹ reviewed their experience with 73 patients with nonfunctioning islet cell carcinoma of the pancreas and concluded that resection of the primary tumor in patients with nonmetastatic disease should be considered. In their series of 33 patients, Legaspi and Brennan¹² demonstrated that the resected group had a projected 3-year survival rate of 100% compared with 34% in patients who had biopsy and chemotherapy. However, Thompson et al.¹⁵ did not find any survival advantage of patients who were deemed to have had a curative resection. We, therefore, are cautious in making any definitive conclusion because our findings may represent selection bias, since it is likely that tumors with a more favorable outcome at the time of presentation are also those that are more amenable to extirpation.

The absence of liver metastases was a significant predictor of survival in several large series.^9,14–16 We also made a similar observation. In their review of 58 patients over a 20-year period, Thompson et al.¹⁵ demonstrated a significant 3-year survival advantage of 82% for patients without liver metastases versus 56% for patients with evidence of liver metastases.

Liver metastases will occur in most patients. In our experience, aggressive treatment of liver metastases with resection, chemoembolization, and/or intra-arterial infusion chemotherapy significantly improved the patient’s chance for long-term survival. Thompson et al.¹⁵ demonstrated symptomatic improvement in more than 50% of patients, with a mean duration of 39 months in patients who underwent noncurative resection. In our series, more than 75% of the patients will have liver involvement during the course of their disease, and more than half will have liver metastases on presentation. Thompson et al.¹⁵ suggested that nonfunctioning tumors tend to be more advanced at the time of diagnosis and attributed this to the absence of significant hormonal production for timely diagnosis. In our experience, this was not the case. The likelihood of having liver metastases at the time of diagnosis was not dependent on tumor function. Fifty-seven percent of functioning tumors and 59% of nonfunctioning tumors presented with synchronous liver metastases.

Selected patients with metastatic neuroendocrine tumors may benefit from liver resections.^17–19 Similar to patients with colorectal metastases, patients with metachronous liver metastases tend to do significantly better than patients with synchronous lesions. Although patients with synchronous liver metastases may have a lower survival from the time of initial diagnosis than patients with metachronous lesions, we do not believe that this should preclude them from undergoing aggressive treatments, either for curative or palliative intent. Although most patients will present with advanced disease, significant long-term survival can still be achieved with surgical and aggressive treatment of the liver metastases.^{12,15,17,18,20} Because of the variables that may affect outcome, the rarity of the disease, and the difficulties associated with controlled clinical trials, any definitive analysis should be viewed with caution. Additionally, one should be cautious when interpreting the results of a therapy because a positive outcome may not necessarily affect the natural history of the disease if there is no control arm for comparison. Despite such difficulties, however, one may still be able to draw some conclusions with regard to the treatment of these patients. Laparoscopic radiofrequency ablation of hepatic neuroendocrine tumor metastases has been reported by Siperstein et al.²¹ and may be beneficial in selected patients.

The 5-year survival for patients with neuroendocrine tumors of the pancreas approaches 35% to 54%.^11,14,15,22 In our series, the 5-year disease-specific actuarial survival was 36%. Although patients with neuroendocrine tumors of the pancreas can potentially be cured, they will require lifelong surveillance because the development of liver metastases can occur 5 to 10 years after treatment of the primary tumor.

	CONCLUSIONS

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
In caring for patients with neuroendocrine tumors of the pancreas, several caveats should be considered: (1) most patients will present with regional or liver metastases, (2) definitive surgical resection of the primary tumor and aggressive treatments of the liver metastases in selected patients should be attempted, (3) patients who do not have liver metastases or who have metachronous liver metastases tend to have a longer survival, and, finally, (4) age, sex, location of the primary tumor, and tumor functional status have no effect on survival.

	Footnotes

 
Presented as a poster presentation at the Society of Surgical Oncology 55th Annual Cancer Symposium, Denver, Colorado, March 14–17, 2002.

An aggressive approach should be considered in selected patients who present with neuroendocrine tumors of the pancreas. Predictive factors of long-term survival (5 years) include definitive resection of the primary tumor, absence of liver metastases, metachronous liver metastases, and aggressive treatment of liver metastases, when present.

Received for publication February 20, 2002. Accepted for publication June 17, 2002.

	REFERENCES

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 

1. Moldrow RE, Connelly RR. Epidemiology of pancreatic cancer in Connecticut. Gastroenterology 1968; 55: 677–86.[Medline]
2. DeVita VT Jr, Hellman S, Rosenberg SA. Cancer Principles and Practice of Oncology. 2nd ed. Philadelphia: JB Lippincott, 1985.
3. Kaplan E, Meier P. Nonparametric estimation from complete observations. J Am Stat Assoc 1958; 53: 457–81.[CrossRef]
4. Cox DR. Regression models and life tables. J R Stat Soc 1972; 34: 187–220.
5. Liu TH, Zhu Y, Cui QC, et al. Nonfunctioning pancreatic endocrine tumors: an immunohistochemical and electron microscopic analysis of 26 cases. Pathol Res Pract 1992; 188: 191–8.[Medline]
6. Delcore R, Friesen SR. Gastrointestinal neuroendocrine tumors. J Am Coll Surg 1994; 178: 187–211.[Medline]
7. Prinz RA, Marangos PJ. Serum neuron-specific enolase: a serum marker for nonfunctioning pancreatic islet cell carcinoma. Am J Surg 1983; 145: 77–81.[Medline]
8. Broughan TA, Leslie JD, Soto JM, Hermann RE. Pancreatic islet cell tumors. Surgery 1986; 99: 671–8.[Medline]
9. Lo CY, van Heerden JA, Thompson G, et al. Islet cell carcinoma of the pancreas. World J Surg 1996; 20: 878–84.[CrossRef][Medline]
10. White TJ, Edney JA, Thompson JS, Karrer FW, Moor BJ. Is there a prognostic difference between functional and nonfunctional islet cell tumors? Am J Surg 1994; 168: 627–30.[Medline]
11. Evans DB, Skibber JM, Lee JE, et al. Nonfunctioning islet cell carcinoma of the pancreas. Surgery 1993; 114: 1175–82.[Medline]
12. Legaspi A, Brennan M. Management of islet cell carcinoma. Surgery 1988; 104: 1018–23.[Medline]
13. Madura JA, Cummings OW, Wiebke EA, Broadie TA, Goulet RL, Howard TJ. Nonfunctioning islet cell tumors of the pancreas: a difficult diagnosis but one worth the effort. Am Surg 1997; 63: 573–8.[Medline]
14. Grama D, Eriksson B, Martensson H, et al. Clinical characteristics, treatment and survival in patients with pancreatic tumors causing hormonal syndromes. World J Surg 1992; 16: 632–9.[CrossRef][Medline]
15. Thompson GB, van Heerden JA, Grant CS, Carney JA, Ilstrup DM. Islet cell carcinomas of the pancreas: a twenty-year experience. Surgery 1988; 104: 1011–7.[Medline]
16. Matthews B, Heniford B, Reardon P, Brunicardi F, Greene F. Surgical experience with nonfunctioning neuroendocrine tumors of the pancreas. Am Surg 2000; 66: 1116–23.[Medline]
17. McEntee G, Nagorney D, Kvols L, Moertel C, Grant C. Cytoreductive hepatic surgery for neuroendocrine tumors. Surgery 1990; 108: 1091–6.[Medline]
18. Norton JA, Sugarbaker PH, Doppmon JL, et al. Aggressive resection of metastatic disease in selected patients with malignant gastrinoma. Ann Surg 1986; 203: 352–9.[Medline]
19. Nave H, Mossinger E, Feist H, Lang H, Raab H. Surgery as primary treatment in patients with liver metastases from carcinoid tumors: a retrospective, unicentric study over 13 years. Surgery 2001; 129: 170–5.[CrossRef][Medline]
20. Eckhauser FE, Cheung PS, Vinik AI, Strodel WE, Lloyd RV, Thompson NW. Nonfunctioning malignant neuroendocrine tumors of the pancreas. Surgery 1986; 100: 978–87.[Medline]
21. Siperstein AE, Rogers SJ, Hansen PD, et al. Laparoscopic thermal ablation of hepatic neuroendocrine tumor metastases. Surgery 1997; 122: 1147–55.[CrossRef][Medline]
22. van Heerden JA. Pancreatic resection for carcinoma of the pancreas: Whipple versus total pancreatectomy—an institutional perspective. World J Surg 1984; 8: 880–8.[CrossRef][Medline]

This article has been cited by other articles:

				S. Ekeblad, B. Skogseid, K. Dunder, K. Oberg, and B. Eriksson Prognostic Factors and Survival in 324 Patients with Pancreatic Endocrine Tumor Treated at a Single Institution Clin. Cancer Res., December 1, 2008; 14(23): 7798 - 7803. [Abstract] [Full Text] [PDF]

				T. R Halfdanarson, J. Rubin, M. B Farnell, C. S Grant, and G. M Petersen Pancreatic endocrine neoplasms: epidemiology and prognosis of pancreatic endocrine tumors Endocr. Relat. Cancer, June 1, 2008; 15(2): 409 - 427. [Abstract] [Full Text] [PDF]

				D. J. Kwekkeboom, W. W. de Herder, B. L. Kam, C. H. van Eijck, M. van Essen, P. P. Kooij, R. A. Feelders, M. O. van Aken, and E. P. Krenning Treatment With the Radiolabeled Somatostatin Analog [177Lu-DOTA0,Tyr3]Octreotate: Toxicity, Efficacy, and Survival J. Clin. Oncol., May 1, 2008; 26(13): 2124 - 2130. [Abstract] [Full Text] [PDF]

				R. Bettini, L. Boninsegna, W. Mantovani, P. Capelli, C. Bassi, P. Pederzoli, G. F. Delle Fave, F. Panzuto, A. Scarpa, and M. Falconi Prognostic factors at diagnosis and value of WHO classification in a mono-institutional series of 180 non-functioning pancreatic endocrine tumours Ann. Onc., May 1, 2008; 19(5): 903 - 908. [Abstract] [Full Text] [PDF]

				M. C. Zatelli, M. Torta, A. Leon, M. R. Ambrosio, M. Gion, P. Tomassetti, F. De Braud, G. Delle Fave, L. Dogliotti, E. C d. Uberti, et al. Chromogranin A as a marker of neuroendocrine neoplasia: an Italian Multicenter Study Endocr. Relat. Cancer, June 1, 2007; 14(2): 473 - 482. [Abstract] [Full Text] [PDF]

				K. K. Kazanjian, H. A. Reber, and O. J. Hines Resection of Pancreatic Neuroendocrine Tumors: Results of 70 Cases Arch Surg, August 1, 2006; 141(8): 765 - 770. [Abstract] [Full Text] [PDF]

				P. Tomassetti, D. Campana, L. Piscitelli, R. Casadei, D. Santini, F. Nori, A. M. Morselli-Labate, R. Pezzilli, and R. Corinaldesi Endocrine pancreatic tumors: factors correlated with survival Ann. Onc., November 1, 2005; 16(11): 1806 - 1810. [Abstract] [Full Text] [PDF]

				M. A. Kouvaraki, J. A. Ajani, P. Hoff, R. Wolff, D. B. Evans, R. Lozano, and J. C. Yao Fluorouracil, Doxorubicin, and Streptozocin in the Treatment of Patients With Locally Advanced and Metastatic Pancreatic Endocrine Carcinomas J. Clin. Oncol., December 1, 2004; 22(23): 4762 - 4771. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Chu, Q. D. Articles by Gibbs, J. F. Search for Related Content PubMed PubMed Citation Articles by Chu, Q. D. Articles by Gibbs, J. F. Related Collections Prognostic factors