Muller, C.O.*; Guérin, F.*; Goldzmidt, D.*; Fouquet, V.*; Franchi-Abella, S.†; Fabre, M.‡; Branchereau, S.*; Martelli, H.*; Gauthier, F.*
Except for surgery of hyperinsulinism during early infancy, indications for major pancreatic resection of tumors or congenital intraparenchymal pancreatic malformations are uncommon during childhood. Therefore, articles regarding pancreatic surgery in children are few, involve a wide range of pancreatic lesions including usually surgery for hyperinsulinism, and comprise mostly case reports or short series without long-term follow-up. Although complete surgical resection is widely admitted to be the only way to improve survival in pancreatic malignancies (1,2), there are few data available in the literature about the long-term pancreatic function and growth of children following pancreatic resections (3), except the study from Shamberger in 1994 (4). The aim of the present study was to report our institutional experience with pancreatic resections focusing on the postoperative complications, and mainly on growth, linked with long-term pancreatic exocrine and endocrine function.
We retrospectively reviewed the charts of all patients younger than 18 years who had undergone pancreatic resection for a pancreatic mass in Bicêtre Hospital between 1986 and 2008. Approval for the study was obtained from our institutional research ethics board. The specific problem of pancreatic resections for hyperinsulinism was excluded from the present study. We have studied for each patient the preoperative clinical, biological, and imaging data, and result of biopsy when performed. We detailed the type of surgical resection and final pathological diagnosis. The postoperative course study included management of abdominal and gastric tubes, complications, length of hospital stay, and parenteral nutrition time. We have collected the present status 2 years after surgery and at last follow-up: weight, height, and evidence or not of diabetes or exocrine insufficiency. Children operated on for malignancies were assessed for survival with/without disease, or death. Follow-up of all but 4 survivors could be updated by means of telephone call during past year.
Eighteen patients (9 girls and 9 boys) with a median age at surgery of 8.9 years (range 0.2–16.2) were identified from charts and office notes.Patients’ data (sex, final diagnosis, age at diagnosis, initial symptoms, presence/absence of metastasis for malignant tumors, type of surgery, postoperative course, and present status) are summarized in Table 1.
Patients’ Histories and Diagnoses
Of the 7 patients with solid and pseudopapillary tumor (SPT), 3 (no. 2,4,5) presented with abdominal pain, related to acute pancreatitis in 1 case; 3 other SPTs (no. 1,3,7) were detected during imaging after abdominal trauma; the last SPT patient (no. 6), a 16-year-old girl, self-discovered a palpable mass of left upper quadrant. Three of these 7 patients had fine-needle biopsy at diagnosis (1 noncontributive), and 1 had an echoendoscopic biopsy. Diagnosis of SPT in the other 3 cases was assessed on the basis of the appearance of the mass on ultrasound (US) and computed tomography (CT) scan, completed by magnetic resonance imaging in 5 cases. None of the 7 patients showed either suspicious regional nodes or distant metastases.
From the 3 patients diagnosed with neuroblastoma (NB), 1 infant (no. 8) presented at the end of the 1980s with an abdominal mass diagnosed, because of an elevated urinary catecholamine level, as a medial nonmetastatic retropancreatic NB; he received preoperative chemotherapy and the accurate diagnosis of intrapancreatic NB was made during surgery. A second primary pancreatic NB (no. 9) was discovered during investigation of a girl presenting with an opsomyoclonic syndrome followed by obstructive jaundice. The third patient (no. 10), a boy, was treated initially by surgery, neoadjuvant chemotherapy, and radiotherapy for a thoracic NB; a CT scan performed during follow-up demonstrated a 3-cm mass developed from the lesser curvature of stomach to the body of the pancreas, which was believed to be a distant metastasis of the thoracic NB. The final pathologic diagnosis was ganglioneuroma invading pancreas.
The first patient with biliary RMS (no. 11) was misdiagnosed and submitted to primary surgery with resection of the bile duct and Roux-en-Y diversion; after postoperative chemotherapy, he was referred to our team for pancreatic resection to remove a tumoral remnant invading the inferior part of the choledocus and the ampulla. In the second patient (no. 12), biliary RMS was diagnosed during postchemotherapeutic follow-up of a scapular RMS and treated by resection of bile duct and Roux-en-Y diversion in another center; this girl was referred to our team for second-look surgery after a relapse in the liver hilum, and then for a second relapse invading pancreas; pancreatic surgery was proposed and performed as a palliative procedure.
The history of patient no. 13 was that of a tumor of the right kidney invading inferior vena cava and retroperitoneal fat. He underwent a right nephrectomy with caval thrombectomy; the first pathological diagnosis was atypical angiomyolipoma, and therefore he received no further treatment. One year later, he developed a huge mass of uncertain diagnosis in the right flank, which was finally reduced after 2 unsuccessful and 1 successful regimens of chemotherapy to a small remnant sticking to both inferior vena cava and the head of the pancreas; therefore, a pancreatic resection associated with a cavectomy was planned to ensure complete excision of tumoral tissues.
The last 2 patients with tumors had a differentiated endocrine carcinoma (DEC, no. 14) and an acinocellular carcinoma (ACC, no. 15), respectively. Both presented with jaundice and were investigated using echoendoscopic biopsy. Both were misdiagnosed, as a pseudoneuroectodermal tumor for the patient with DEC and as an atypical SPT for the patient with ACC. In the latter patient, peritoneal metastases were found at surgery; nevertheless, pancreatic resection was possible and achieved to release jaundice.
The first patient (no. 16) with cystic malformation of the pancreas benefited from prophylactic surgery after antenatal diagnosis. The second one (no. 17) had curative surgery after a 6-month history of recurrent episodes of acute pancreatitis with paraesophageal and pleural effusion despite an unsuccessful attempt to drainage of a supposed pseudocyst.
The diagnosis of chronic calcifying pancreatitis predominating in the tail of the gland was established in the last patient of the series (no. 18) after a 1-year history of recurrent acute pancreatitis episodes.
Surgery and Short-term Postoperative Course
Eleven pancreaticoduodenectomies (PDDs) were performed for patients with cephalic lesions, 2 splenopancreatectomies (SPs) for patients with caudal lesions involving the splenic hilum, 2 mid-pancreatic resections (MPRs) for patients with isthmic lesions, 2 tumorectomies for patients with peripheral lesions, and 1 caudal pancreatectomy with Roux-en-Y diversion for the patient with chronic calcifying pancreatitis. Regarding the 11 PDDs, the pylorus was preserved in 8 cases requiring no extensive lymphadenectomy, and was not preserved in 3 cases requiring large margins (Figs. 1 and 2).
The median operating time was 6.5 hours (range 4–10) for PDD, 5.25 hours for SP, 5 hours for MPR, and 2 hours for tumorectomies. Transfusion was needed in 3 cases during PDD: two 500 mL and one 1000 mL.
Concerning the 15 patients with tumors, complete microscopic resection could be achieved in only 11 patients. In patients 12 (recurrent RMS) and 13 (acinar cell carcinoma with peritoneal carcinomatosis not seen on CT scan), PDDs were performed as palliative procedures. In patients 9 and 10 (NB), macroscopic resection with positive margins could be achieved; both patients are in complete remission 3 and 20 years after surgery, respectively.
Postoperative mortality was null. The major complication related to surgery was associated pancreatic and biliary fistulas observed after a palliative atypical PDD in 1 patient (no. 12), and on the cephalic side of an MPR in a second patient (no. 2) who required reoperation for Roux-en-Y biliary diversion and pancreatic drainage; the overall fistula rate after PDD is thus 9%. Minor complications were observed after PDD: delayed gastric emptying in 2 patients, transient pleural effusion in 1, intussusception in 1, and transient dilatation of the pancreatic duct in 1 patient. Another major complication occurred in one of the patients referred to PDD after a history of recurrent acute pancreatitis episodes with abdominothoracic pseudocyst (no. 15): a portal thrombosis was detected by US monitoring at day 7; it was treated by percutaneous thrombectomy, followed by 2 balloon dilatations and finally placement of a stent. Because the stenosis was located at the upper border of the Winslow foramen, it could be regarded as a sequalae of the repeated episodes of pancreatitis or as a postoperative complication; the question is pending.
The median duration of total or partial parenteral nutrition is 14 days in PDD (range 0–60), 12 days in MPR (range 0–23), 8 days in distal pancreatectomy, and was not required in SP and tumorectomy. The median time for nasogastric tube removal was 4.5 days (range 1–20), and median time for abdominal drainage removal was 8.5 days (range 4–46). The median postoperative hospital stay was 15 days (range 6–22) in an uneventful postoperative course and 24 days (range 14–53) in complicated cases.
The mean follow-up time was 4.2 years, with a range of 2 to 11 years. Among the patients with tumors, 3 with cancer (2 patients with RMS and 1 with ACC) died of spread of uncontrolled disease. One patient with NB treated by tumorectomy (no. 8) was lost to follow-up after 12 months because the family moved; he was tumor free at that time. The remaining 11 children are alive and disease free (7 SPT, 3 NB, 1 DEC). The last 3 patients with benign lesions are alive and well.
None of the patients experienced diabetes mellitus. One patient (no. 2 with MPR for SPT complicated by a pancreatic and biliary fistula) developed fat intolerance with long-term need for pancreatic enzyme therapy. The BMI of the 3 patients (no. 11, 12, 15) who died of malignancies decreased as soon as their disease became uncontrolled.
Ten of the 15 survivors had steady growth and BMI curves in the normal range, within the 5th to 95th and the 3rd to 97th percentiles, respectively. Three teenage girls (no. 4, 6, 14) moved from upper to mean normal BMI values in a 2- to 4-year period. In a boy (no. 1) with a known risk behavior, the BMI decreased from the 85th to the 3rd percentile during a 4-year period, without any evidence of diabetes or exocrine pancreatic insufficiency. The last patient (no. 5) was obese before surgery and was still obese 3 years later (age 16 years).
The question of the management of a pancreatic mass may be discussed in terms of indications, choice of surgical techniques, and follow-up. Although the major part of articles about pancreatic surgery comes from adult experience, some data regarding children are available.
According to Japanese autopsy data (5), pancreatic cancers account for <0.2% of deaths from malignant disease in children, and Jaksic et al (6) found in 1992 that an operation for a newly diagnosed pancreatic cancer occurs only once every 18,000 pediatric and adult surgical cases. Recently, Perez et al (2) found that the overall incidence of malignant pancreatic tumors in children was 0.018 newly diagnosed cases per 100,000 people in the SEER population (Surveillance, Epidemiology, and End Results registry, Miami, FL), this number appearing to increase from 1979 to 2002. No registry detailing the overall relative incidence of the different malignant and benign pancreatic tumors in children is available. In our series, among 15 tumors, 8 were epithelial (7 SPT and 1 ACC) and 7 were nonepithelial (3 NB, 2 biliary RMS, 1 DEC, and 1 extension of RAA), and there was no pancreatoblastoma. Pooling data from 4 American and 1 Canadian series with a total number of 94 patients (1,2,6,7), these tumors appear with a different histological spectrum, are predominant in girls, and are associated with a better prognosis compared with the adults. These features result from the fact that the most frequent tumor in childhood is the solid and pseudopapillary tumor, which counts for 2% to 3% of all pancreatic tumors (8), is predominant in girls (male vs female 1:2.5 in our series) during the second decade of life (median age 13.2 years in our series), and is a low-grade malignant tumor. We reported no synchronous metastasis nor recurrence in our series, whereas Choi et al (8) reported 1 case of metachronous multiple liver metastasis in 23 patients and Lee et al (9) reported 1 case of synchronous liver metastasis and 1 case of distant (7 years) peritoneal recurrence in 15 patients. Primary complete surgical resection of nonmetastatic SPT is mandatory to ensure an expected survival rate without recurrence of 100% in our series, as in others (8,9). This is evenly justified by the lack of data about response to chemotherapy of metastatic SPT (1). Recent case reports mention the possibility of neoadjuvant chemotherapy by gemcitabine for downsizing before definitive surgical resection (10,11).
In adults with a predominance of ductal adenocarcinoma, jaundice and weight loss are the most common presentations in 72% and 45% of cases, respectively (12). Conversely, initial clinical signs associated with pancreatic masses are various in children. We report abdominal trauma (16%), palpable mass (16%), jaundice (16%), and abdominal pain (11%). This heterogeneity of presentations in childhood may be explained by the variety of histological types of tumors acting usually more as “pushers” than as “invaders.” US examination demonstrates the pancreatic origin of the tumor. CT and magnetic resonance imaging examination can be typical in some cases of pseudopapillary tumors, for example, a large tumor encapsulated, with marked degenerative and hemorrhagic changes, with or without calcifications in the capsule (13). The large spectrum of diagnosis associated with unspecific clinical signs, however, sometimes makes a preoperative biopsy preferable before planning radical surgery in children. Moreover, pancreatic masses happened to be chronic pancreatitis in adults and also in children (14). In fact, 6 fine-needle aspiration biopsies had been performed in our series, 3 transparietal and 3 by echoendoscopy. Because of the small number of patients, the sensitivity is approximately 50% and the morbidity is 16%: fistula of the dilated common hepatic duct treated by nasobiliary drainage for 1 month in the patient with NB invading the head of the pancreas. Histological diagnosis may be difficult, as proved by the 2 discrepancies in our series: 1 DEC misdiagnosed as a PNET and 1 ACC misdiagnosed as an atypical SPT. In adults, radiologists and echoendoscopists have a good experience in pancreatic biopsy, showing a sensitivity of 68% in cytologic and histological examination and an extremely low morbidity around 1.6% (15).
Positive indications for pancreatic surgery are primary tumors of the pancreas and congenital pseudocysts. All of the studies, despite the small number of patients, conclude that early complete and even aggressive surgical resection is the key to survival improvement for people with pancreatic malignancies (1,2,6,7). The nontraumatic pseudocysts do not regress spontaneously (16,17); this—in the absence of another identified cause—favors a congenital origin. The choice of the type of resection must take into account several variables: the location of the mass within the pancreas, its contacts with portal vein and mesenteric artery and with the surrounding organs, and the potential malignancy of the tumors, which need at least complete macro- and microscopic resection, and sometimes lymph nodes picking.
To achieve macroscopic resection while avoiding complications and sparing endocrine pancreas, the masses located in the head of the pancreas must be treated by PDD with preservation of the pylorus or the Whipple procedure (8/11) if benign or of low-grade malignancy, or without preservation of the pylorus if high-grade malignancy (3/11). These results are similar to the adult literature, approximately 80% of the Whipple procedure (15,18). No data are available in the pediatric literature about biliodigestive diversion and risk of cancer: 1 Italian study in 2001 mentions an incidence of cholangiocarcinoma of 1.9% after hepaticojejunostomy in a series of 1003 patients from 1967 to 1997; the presence of cholangitis was the unique factor affecting the incidence of cholangiocarcinoma in univariate and multivariate analyses (19). Central tumors can be treated by MPR, although this surgery presents an important risk of complication such as pancreatic and biliary fistula. Tumors located in the tail of the pancreas can be removed easily by caudal pancreatectomy. Associated splenectomy must be restricted to tumors invading the splenic hilum as in our series because of the short- and long-term risks of splenectomy: infection and long-term adverse effects of thrombocythemia, such as pulmonary hypertension.
Short- and long-term morbidity of pancreatic surgery has been widely studied in the adult literature, based on huge cohorts of patients; however, there are few data available about short- and long-term outcomes of this surgery in children. First, pediatric surgeons are not familiar with the PDD, and it is widely admitted that the experience of the surgical center has a close relation to better clinical outcomes in adult pancreatic surgery. Dasgupta and Kim (18) demonstrated that there was minimal difference between adult and pediatric duodenopancreatectomy, in terms of operating time, admission time, and mortality. We showed an average operating time of 366 minutes (range 240–600) for PDD, which compares favorably with the adult literature: 420 minutes, and even to the Cameron study results, 330 minutes (15). Postoperative mortality usually hovers near 1%, whereas we had no deaths related to the surgery. A high rate of short-term complications (45%), as in adults, was observed after PDD in our series, but included only 1 case of fistula requiring a prolonged hospital stay and no infection. Still, it is always difficult to compare the adult and the pediatric literature because adult patients who undergo pancreatic resection are often elderly, with multiple debilitating conditions (18). The median admission time is long: 19 days (vs 9 days in the Cameron et al study), but with no loss of weight during admission, which is certainly the result of almost systematic parenteral nutrition in PDD.
More important, with an average follow-up of 4.2 years and a range of 2 to 11 years, we reported no cases of diabetes mellitus independent of the type of surgery, and only 1 case of fat diet intolerance after an MPR but none after PDD. In our experience, anomalies of growth and BMI evolution after pancreatic resection were observed only in patients with uncontrolled malignancies or with unrelated risk factors (risk behavior or morbid obesity). This fact suggests that partial pancreatic resections, including PDD, do not impair growth in childhood.
Because of the great variety of diseases resulting in pancreatic solid and cystic masses, an accurate diagnosis based on histological proof in tumors is mandatory before planning aggressive surgery. It is widely admitted that, provided a complete resection, the prognosis is globally better for children having a pancreatic neoplasm than for adults when operated upon. Therefore, pediatric surgeons must not be afraid of pancreatic surgery, even PDD, which has in childhood an extremely low mortality, a short-term morbidity similar to that observed in adults, and a low risk of endocrine or exocrine pancreatic dysfunction, and does not seem to impair growth in younger children.
1. Shorter NA, Glick RD, Klimstra DS, et al. Malignant pancreatic tumors in childhood and adolescence: the Memorial Sloan-Kettering experience, 1967 to present. J Pediatr Surg
2. Perez EA, Gutierrez JC, Koniaris LG, et al. Malignant pancreatic tumors: incidence and outcome in 58 pediatric patients. J Pediatr Surg
3. Rabinovich A, Rescorla FJ, Howard TJ, et al. Pancreatic disorders in children: relationship of postoperative morbidity and the indication for surgery. Am Surg
4. Schamberger RC, Hendren WH, Leichtner AM. Long-term nutritional and metabolic consequences of pancreaticoduodenectomy in children. Surgery
5. Sukimoto I, Watanabe K, Lin J, et al. Pancreatic carcinoma in children in Japan. Cancer
6. Jaksic T, Yaman M, Thorner P, et al. A 20 year review of pediatric pancreatic tumors. J Pediatr Surg
7. Grosfeld JL, Vane DW, Rescorla FJ, et al. Pancreatic tumors in childhood: analysis of 13 cases. J Pediatr Surg
8. Choi SH, Kim SM, Oh JT, et al. Solid pseudopapillary tumor of the pancreas: a multicenter study of 23 pediatric cases. J Pediatr Surg
9. Lee SE, Jang JY, Hwang DW, et al. Clinical features and outcome of solid pseudopapillary neoplasm: differences between adults and children. Arch Surg
10. Maffuz A, Bustamante FT, Silva JA, et al. Preoperative gemcitabine for unresectable, solid pseudopapillary tumour of the pancreas. Lancet Oncol
11. Kanter J, Wilson DB, Strasberg S. Downsizing to resectability of a large solid and cystic papillary tumor of the pancreas by single-agent chemotherapy. J Pediatr Surg
12. Cameron JL, Riall TS, Coleman J, et al. One thousand consecutive pancreaticoduodenectomies. Ann Surg
13. Chung EM, Travis MD, Conran RM. Pancreatic tumors in children: radiologic-pathologic correlation. Radiographics
14. Park J, Dunn JC, Atkinson JB. Management of children with pancreatic head mass. J Pediatr Surg
15. Iglesias-Garcia J, Dominguez-Munoz E, Lozano-Leon A, et al. Impact of endoscopic ultrasound-guided fine needle biopsy for diagnosis of pancreatic masses. World J Gastroenterol
16. Stone HH, Whitehurst JO. Pseudocysts of the pancreas in children. Am J Surg
17. Teha SH, Phama TH, Leeb A, et al. Pancreatic pseudocyst in children: the impact of management strategies on outcome. J Pediatr Surg
18. Dasgupta R, Kim PCW. Relationship between surgical volume and clinical outcome: should pediatric surgeons be doing pancreaticoduodenectomies? J Pediatr Surg
19. Tocchi A, Mazzoni G, Liotta G, et al. Late development of bile duct cancer in patients who had biliary-enteric drainage for benign disease: a follow-up study of more than 1,000 patients. Ann Surg