SEARCH

SEARCH BY CITATION

Keywords:

  • Wilms tumor;
  • surveillance imaging;
  • computed tomography;
  • pediatrics;
  • tumor relapse

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

BACKGROUND:

It is unclear whether routine pelvic imaging is needed in patients with Wilms tumor. Thus, the primary objective of the current study was to examine the role of routine pelvic computed tomography (CT) in a cohort of pediatric patients with Wilms tumor.

METHODS:

With institutional review board approval, the authors retrospectively identified 110 patients who had Wilms tumor diagnosed between January 1999 and December 2009 with surveillance imaging that continued through March 2011. The authors estimated overall survival (OS), event-free survival (EFS), and dosimetry from dose length product (DLP) conversion to the effective dose (ED) for every CT in a subgroup of 80 patients who had CT studies obtained using contemporary scanners (2002-2011). Metal-oxide-semiconductor field-effect transistor (MOSFET) dosimeters were placed within organs of anthropomorphic phantoms to directly calculate the truncal ED. EDDLP was correlated with EDMOSFET to calculate potential pelvic dose savings.

RESULTS:

Eighty patients underwent 605 CT examinations that contained DLP information, including 352 CT scans of the chest, abdomen, and pelvis; 123 CT scans of the chest and abdomen; 102 CT scans of the chest only; 18 CT scans of the abdomen and pelvis; 9 CT scans of the abdomen only; and 1 CT that was limited to the pelvis. The respective 5-year OS and EFS estimates were 92.8% ± 3% and 2.6% ± 4.3%. Sixteen of 110 patients (15%) developed a relapse a median of 11.3 months (range, 5.0 months to 7.3 years) after diagnosis, and 4 patients died of disease recurrence. Three patients developed pelvic relapses, all 3 of which were symptomatic. The estimated ED savings from sex-neutral CT surveillance performed at a 120-kilovolt peak without pelvic imaging was calculated as 30.5% for the average patient aged 1 year, 30.4% for the average patient aged 5 years, 39.4% for the average patient aged 10 years, and 44.9% for the average patient aged 15 years.

CONCLUSIONS:

Omitting pelvic CT from the routine, off-therapy follow-up of patients with Wilms tumor saved an average 30% to 45% of the ED without compromising disease detection. Cancer 2013. © 2012 American Cancer Society.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

Wilms tumor is the most common renal tumor in children and accounts for approximately 6% of all malignancies in children aged <15 years.1 With current therapies, greater than 90% of children are expected to be cured of their disease, and a significant proportion of patients who fail primary therapy can be salvaged.2 Given the high cure rates and favorable outcomes after relapse, the usefulness of routine comprehensive imaging monitoring is controversial.3-6 Imaging-associated patient exposures to ionizing radiation are of increasing concern7-11 and may be considerable, particularly in the growing population of young patients who undergo repeated imaging related to cancer treatment and follow-up.

The lung is the most common site of relapse for Wilms tumor.12-14 Abdominal and pelvic relapses are rare, occurring in approximately 10% of patients.12, 13 Despite this, pelvic surveillance imaging is recommended in many centers and in the current front-line Wilms tumor trials of the Children's Oncology Group. It is unclear whether the early detection of abdominopelvic relapse has an impact on clinical outcomes, because approximately 50% of patients13 who relapse after treatment with vincristine, actinomycin-D, and doxorubicin plus radiation therapy12 and approximately 80% of those who relapse after initial treatment with vincristine and actinomycin-D14 are expected to survive long term. Few studies have assessed the impact of surveillance imaging on the detection of Wilms tumor recurrence, salvage rates at the time of relapse, and long-term survival outcomes. Thus, the primary objectives of the current study were to examine the role of routine pelvic computed tomography (CT) surveillance in pediatric patients with Wilms tumor and to examine the rate of pelvic relapse and long-term survival in a population of patients with Wilms tumor treated at our center. Our secondary objectives were to estimate the associated patient exposure to ionizing radiation from CT and to investigate the potential radiation dose savings when excluding pelvic imaging for follow-up care.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

Patient Inclusion/Exclusion for This Study

With institutional review board approval, we retrospectively identified patients who had Wilms tumor initially diagnosed at our institution between January 1999 and December 2009 and who received surveillance imaging through March 2011. For this patient group, we analyzed outcome and recorded patient demographics, disease stage at diagnosis, treatment regimen, disease relapse and location, the number of CT scans each patient underwent, and further delineated CT scans that included the pelvis. Tumors were staged according to the National Wilms Tumor Study (NWTS) Group surgical-pathologic staging system.15 Relapses were coded as local, regional, distant, and combined. Metachronous occurrence of Wilms tumor was coded as locoregional relapse. All data were managed in compliance with the Health Insurance Portability and Accountability Act of 2006.

For patient dosimetry, we analyzed 80 patients (of the 110) who had 605 CT examinations performed between 2002 and 2011 (n = 80 patients; 605 examinations). CT studies performed before 2002 (n = 23 patients) were excluded from dose estimations, because no patient dosimetry data (ie, dose length product [DLP]) were recorded, and CT technology changed from single-detector CT to contemporary multidetector CT. In addition, 7 patients from 1999 to 2011 were excluded from dose calculations because they had no CT studies of the chest, abdomen, or pelvis available for review. Abdominal studies were defined as scans from the diaphragm to the iliac crest, and pelvic studies were defined as scans from the iliac crest to the perineum (inferior aspect of the symphysis pubis).

Computed Tomography Technique

Patients were scanned on 1 of 4 CT scanner models from 1999 through 2011: the Siemens Plus 4 single detector (1999-2002; Siemens Medical Solutions, Erlanger, Germany), the GE Lightspeed Ultra 8 detector (2002-2007; General Electric Medical Systems, Milwaukee, Wis), the GE Discovery 4 detector positron emission tomography (PET)/CT scanner (2002-2011; General Electric Medical Systems), and the GE VCTXT 64 detector (2008-2011; General Electric Medical Systems). Before 2008, we used a 300-mg/mL concentration of iodine as nonionic, intravenous contrast; and, since 2008 to the present, we have used a 270-mg/mL concentration of iodine as nonionic, intravenous contrast. CT examinations were performed after the administration of oral contrast medium and up to 2.0 mL/kg of intravenous contrast.

Patient Dosimetry

Patient dosimetry estimates were based on the recorded DLP from every CT examination performed using a conversion from DLP to effective dose (ED) calculation methodology.16-18 Briefly, the ED conversion from DLP was calculated by multiplying the patient-specific DLP with an age and region of body weighted conversion factor (k)17-19 as follows:

  • equation image(1)

The DLP for each patient's CT scan obtained over the course of their surveillance was recorded (ie, as a function of age). DLP values for each patient were grouped by integer age (eg, 1 year, 2 years, etc) and were multiplied by the age-appropriate k factor. The DLP-to-ED conversion factors listed in American Association of Physicists in Medicine (AAPM) Report 9617 and recently updated by Huda et al19 were provided only for the groups aged 1 year, 5 years, 10 years, and adults. To account for other age-specific conversion factors not listed in AAPM Report 96, the original, discrete values were interpolated using a cubic function to create a look-up table of continuous data. A software graphic user interface was developed (Fig. 1) to generate age-specific conversion factors and to calculate the ED from DLP data.

thumbnail image

Figure 1. Conversion factors (k) that were used to convert patient-specific dose length product (DLP) values were published only for discrete ages of birth 0, 1 year, 5 years, 10 years, and adults. (Top) Because the patients with Wilms tumor in this study were imaged at all ages, a program to interpolate k values was created based on American Association of Physicists in Medicine (AAPM) Report 9618 and data from Huda et al19 (Bottom) The resulting conversion factor interpolation (lines) is demonstrated for effective dose (ED0 conversions of computed tomography (CT) scans to the trunk region. Original data from AAPM Report 96 as corrected by Huda et al are indicated by circles. Abd indicates abdomen; Pel, pelvis.

Download figure to PowerPoint

Because of the limited number of pelvis-only CT examinations from 2002 to 2011 (n = 1), we could not differentiate the dose to patients from pelvic CT scans using the EDDLP methodology. To estimate dose savings by eliminating the pelvic scan during traditional chest-abdomen-pelvis (CAP) surveillance imaging, metal-oxide-semiconductor field-effect transistor (MOSFET) dosimeters (mobile MOSFET; Best Medical, Ottawa, Ontario, Canada) were placed within organs of anthropomorphic phantoms (CIRS, Norfolk, Va) (Fig. 2) to directly calculate truncal ED. The MOSFET dosimeters were placed in all major organs from lungs through pelvic floor, including breast tissue, lungs, liver, stomach, pancreas, kidneys, intestine, bladder, ovaries/testes, bone surfaces, and bone marrow (ribs, lumbar vertebra, and pelvis). Larger organs (eg, lungs and liver) had 5 or 6 dosimeters placed throughout their volume; all other organs had 2 or 3 dosimeters. MOSFETs were placed in 4 size-adapted phantoms, which were used to emulate various sizes of pediatric body habitus. The phantoms consisted of 5 simulated tissue types: soft tissue, lung, brain, spinal cord, and bone. Tissue types were engineered to produce photon attenuation values within 1% for bone and soft tissue substitute and within 3% for lung tissue substitute over the range of from 30 to 20,000 kiloelectron volts (as stated in the CIRS product literature). MOSFET-loaded phantoms were scanned with CT using Wilms tumor patient-specific CT imaging protocols. Each phantom was scanned a minimum of 3 times, and the recorded MOSFET measurements were averaged. The averaged MOSFET absorbed dose data (DT,R) were multiplied by an organ and tissue weighing factor (WT) specified by International Commission on Radiological Protection publication 10320 and by a radiation weighting factor (WR)16:

  • equation image(2)

where WR for x-rays is equal to unity.21

thumbnail image

Figure 2. Anthropomorphic phantoms provide a more patient-specific means of measuring absorbed dose from CT scanning. (Left) Metal-oxide-semiconductor field-effect transistor (MOSFET) dosimeters were placed within the organs of 4 phantoms ranging from lung to pelvic floor. (Right) The 4 phantoms approximate the 1-year, 5-years, 10-years, and adult age groups represented in American Association of Physicists in Medicine Report 96.18

Download figure to PowerPoint

Statistical Methods

Survival was calculated from the date of diagnosis to the date of death from any cause or the date of last contact for survivors. Event-free survival (EFS) was calculated from the date of diagnosis to the date of the first event (disease relapse, second malignancy, or death from any cause) or the date of last contact for patients without events. The Kaplan-Meier method was used to estimate survival and EFS. Outcome estimates are reported with ±1 standard error. Statistical analyses were performed using the SAS statistical software package (version 9.2; SAS Institute, Inc., Cary, NC).

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

Patient Cohort

We identified 110 patients with newly diagnosed Wilms tumor. Sixty of those 110 patients (55%) were female, and 68 patients (62%) were white. The median age at the time of primary diagnosis was 35 months (range, 3.6 months to 19.9 years). Disease stage at diagnosis was stage I in 20 patients, stage II in 33 patients, stage III in 26 patients, stage IV in 20 patients, and stage V in 11 patients. The median patient age at last follow-up was 9.9 years (range, 1.5-28.0 years).

Treatment

The majority of patients (96%) received risk-directed therapies based on NWTS 5 and included vincristine and actinomycin-D with or without doxorubicin and with or without radiotherapy for patients who had tumors with favorable histology.22 The remaining 4 patients received a regimen that included ifosfamide, carboplatin, etoposide, vincristine, doxorubicin, and cyclophosphamide.

Patient Outcome

Overall, 103 patients (93.6%) remained alive at a median follow-up after diagnosis of 6.5 years (range, 0.1-12.5 years). Almost 90% of survivors (90 of 103 patients; 87%) had been seen or contacted within the last year. Five-year estimates of survival and EFS for all patients were 92.8% ± 3% and 82.6% ± 4.3%, respectively. Sixteen of 110 patients (15%) relapsed at a median of 11.3 months (range, 5.0 months to 7.3 years) after diagnosis, and 4 died from disease recurrence. Two additional patients died from second malignant neoplasms (1 each from medulloblastoma and parameningeal rhabdomyosarcoma), and 1 patient died from a complication of Williams syndrome. Four patients died of relapsed Wilms tumor a median of 10.7 months after the time of relapse (range, 5.7-14.7 months) and a median of 18.9 months after initial diagnosis (range, 17.7-24.7 months). Two patients died from intrathoracic relapse only, 1 died from abdominal relapse only, and 1 died from a combined intrathoracic and abdominal relapse. One patient developed a desmoid tumor of the abdomen, which was surgically resected; this patient remained alive 7.5 years after the time of resection. Sites of relapse or second Wilms tumor occurrence included lungs in 7 patients (associated with hilar adenopathy in 1 patient), the abdomen in 5 patients (including 2 patients who developed metachronous Wilms tumor in the opposite kidney), the pelvis in 3 patients, and a combined site in the abdomen and lung in 1 patient. Three of 110 patients (3%) developed pelvic relapse at 22.0 months, 34.7 months, and 38.8 months after diagnosis; no other sites of disease were identified at the time of pelvic relapse. All patients who had a pelvic relapse were symptomatic at the time of recurrence (1 had painless vaginal bleeding, 1 had urinary retention, and 1 had abdominal pain and constipation). At the time of initial diagnosis, these 3 patients had NWTS stage II, III, and IV disease (Table 1). The 3 patients who had pelvic relapses were alive without evidence of disease 9.0 years, 4.6 years, and 4.2 years after relapse.

Table 1. Characteristics of the 3 Patients Who Developed Pelvic Relapse
PatientSexRaceAge at Diagnosis, yNWTS Disease StageLocal StageHistologySymptoms at Pelvic RelapseTime From Diagnosis to Pelvic Relapse, yTime from Relapse to Last Follow-Up, yDisease Status
  1. Abbreviations: NED, no evidence of disease; NWTS, National Wilms' Tumor Study.

1FemaleBlack4.5IITumor rupture at surgeryFavorableUrinary retention1.89.0Alive NED
2FemaleWhite10.5IVTumor margins <1 mm, positive lymph nodesFavorable; blastemal predominantPainless vaginal bleeding3.24.6Alive NED
3MaleWhite2.6IIIPelvic and scrotal disease at diagnosisFavorableAbdominal pain and constipation2.94.2Alive NED

Computed Tomography Studies

In total, 605 CT scans were performed on 80 patients between 2002 and 2011, including 352 CT scans of the chest, abdomen, and pelvis combined (CAP); 123 CT scans of the chest and abdomen combined; 102 CT scans of the chest only; 18 CT scans of the abdomen and pelvis combined; 9 CT scans of the abdomen only; and 1 CT scan of the pelvis only. Figure 3 illustrates the resulting EDDLP as a function of median patient age at the time of CT examinations. From the time of diagnosis to the date of last follow-up, in total, 371 CT scans included the pelvis (average per patient, 6 scans; range, 1-29 scans). Table 2 provides data on the average patient EDMOSFET and EDDLP estimates for CAP CT scans as a function of approximate anthropomorphic phantom age grouping. From the EDMOSFET calculations, the pelvic ED was isolated, and the estimated ED savings from a sex-neutral CT surveillance scan at a 120-kilovolt peak performed without imaging of the pelvis was calculated at 30.5% for the average patient aged 1 year, 30.4% for the average patient aged 5 years, 39.4% for the average patient aged 10 years, and 44.9% for the average patient aged 15 years. Corresponding decreases to the lifetime risk of cancer incidence and mortality were estimated excluding pelvic CT scans (Table 3).

thumbnail image

Figure 3. Dose length product (DLP) conversion to the effective dose (ED) (EDDLP) is illustrated. Pelvic-only dosimetry is not illustrated, because so few computed tomography examinations of those respective studies were performed on patients with Wilms tumor between 2002 and 2011. mSv, millisievert; CAP, chest-abdomen-pelvis; Abd, abdomen.

Download figure to PowerPoint

Table 2. Comparison of Effective Dose Calculations as Measured by Metal-Oxide-Semiconductor Field-Effect Transistor and Dose Length Product Methodology
 Average±SD, mSv 
Approximate Age, yEDMOSFETEDDLPDifference, %
  1. Abbreviations: DLP, dose length product; ED, effective dose; MOSEFT, metal-oxide-semiconductor field-effect transistor; mSv, millisievert; SD, standard deviation.

13.00.13.60.817
53.20.23.21.01
104.10.24.22.83
158.80.415.50.143
Table 3. Decreased Lifetime Risk for Cancer Incidence and Mortality When Performing a Chest/Abdomen Computed Tomography (CT) Scan Compared With a Chest/Abdomen/Pelvis CT Scan Based on the Scanning Techniques in the Current Work
 Incidence/Mortality, %
 MalesFemales
Approximate Age, yAverage: 6 CT ScansMaximum: 29 CT ScansAverage: 6 CT ScansMaximum: 29 CT Scans
10.1/00.3/00.1/0.10.6/0.3
50.1/0.10.3/0.30.2/0.10.9/0.3
100.3/0.21.5/0.90.2/0.10.9/0.3
150.2/0.10.9/0.60.4/0.11.7/0.6

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

Because approximately 90% of children with Wilms tumor can be cured with contemporary therapy13 and a significant proportion of those who relapse can be salvaged,12-14 the utility of routine, off-therapy surveillance imaging is controversial.3-6, 23 It is also unclear whether the use of imaging to detect clinically asymptomatic disease recurrence in the abdomen and pelvis has an impact on salvage rates or patient outcomes. Because the lungs are the most common site of recurrence, and abdominal and pelvic relapses are reported in only approximately 10% of patients,13, 14 the value of routine abdominal-pelvic surveillance imaging should be questioned.

Kan et al failed to identify any pelvic relapses on routine surveillance scans in a small study that included 16 patients with hepatoblastoma and 17 patients with Wilms tumor. The use of pelvic CT did not modify therapy in these patients. However, that study lacked radiation dosimetry data, and the number of patients was small.24 In a series of 1466 patients with nonmetastatic, favorable histology Wilms tumor, Breslow et al associated the following factors with an increased risk of abdominal relapse: older age at diagnosis (aged ≥48 months), higher stage disease (stage III), specimen weight >1000 g, positive tumor margin, and capsular penetration.14 In another series, factors that contributed the greatest risk of local recurrence were tumor spillage, stage III disease, and unfavorable histology.25

Each of the 3 patients in our study who developed pelvic relapse was symptomatic, recurred within 3 years from diagnosis, and had at least 1 risk factor as defined by Breslow, including age >48 months (n = 2) and higher disease stage (n = 2).14 All 3 patients were salvaged and survived disease free 4.2 to 9.0 years after relapse. Our findings suggest that asymptomatic pelvic relapse is uncommon and rarely occurs 3 years after diagnosis. Little information is available delineating the salvage rate of patients with pelvic disease relapse. Shamberger et al reported that the greatest risk of local disease recurrence was tumor spillage, particularly among patients with stage II disease who received treatment with only 2 drugs. The reported survival rate at 2 years was 43% and was worse for patients who had stage III disease and tumor spillage at the time of initial diagnosis.25 Malogolowkin et al reported that approximately 50% of patients who relapsed after initially receiving vincristine, actinomycin, and doxorubicin could be cured of their disease and documented an improved outcome for patients who presented with ≤3 adverse prognostic signs (stage IV disease, unfavorable histology, relapse within 1 year from diagnosis, ≥2 relapses, thoracic and abdominal relapse, brain or bone metastases, tumor-positive lymph nodes, and postradiation in-field relapse).13 Green et al reported 4-year EFS and overall survival rates of 71% and 82%, respectively, in patients who relapsed after initial therapy with vincristine and actinomycin-D.12 A retrospective review of larger numbers of patients enrolled in Children's Oncology Group trials could help clarify the incidence and outcome of patients with pelvic relapses.

The 80 patients we used to estimate patient dosimetry underwent a collective 605 CT surveillance examinations. No attempt was made to differentiate between initial diagnostic and ongoing surveillance CT examinations of the pelvis. Recently, it has been demonstrated that early childhood exposure to ionizing radiation (aged <20 years) increases the risk of developing solid tumors later in life.26, 27 It has become widely accepted that children and adolescent patients represent a patient population at higher risk from exposure to ionizing radiation from CT for all types of cancer.7 In keeping with “As Low As Reasonably Achievable Radiation” (ALARA) principles, we recommend reconsidering surveillance methodologies by omitting imaging of the pelvis altogether. Should pelvic imaging be considered clinically necessary, then imaging with ultrasound or magnetic resonance (MR) techniques to avoid exposure to ionizing radiation should be chosen while recognizing the strengths and limitations of both modalities.

Neither ultrasound nor MR exposes patients to ionizing radiation. Ultrasound is readily available, requires little or no patient preparation, and is less costly than either CT or MR. However, the quality of the imaging study depends heavily on technologist skill, patient cooperation, and patient body habitus. MR is more costly than either CT or ultrasound; provides detailed, multiplanar information; but may require patient sedation or anesthesia and gadolinium contrast administration, which are not without risk. The development of faster and more specific MR sequences ultimately may obviate the need for sedation and allow more liberal use of MR for disease staging and monitoring.

Our findings must be considered in the context of the limitations of a retrospective review in which treatment eras, follow-up imaging, and imaging techniques evolved over the course of 1 decade. Despite these limitations, we report a modest sized patient population for whom treatment and imaging all were performed at a single institution and for whom cumulative long-term follow-up in excess of 700 patient-years was available.

True patient-specific dosimetry data are not available for multidetector CT. Therefore, patient dosimetry was estimated using the latest techniques reported in the literature. The dosimetry methods described in this work have several limitations. All EDDLP values were derived from cylindrical Plexiglas phantoms, which were indiscriminate with regard to sex and age. In addition, nearly all recorded DLP values were derived from cylindrical phantoms that measured 32 cm in greatest dimension, a size that tends to underestimate exposure risk to smaller pediatric patients by upward of 50%. In this work, we associated EDDLP calculations with EDMOSFET measurements to more specifically reflect differences in radiation risk for weight, age, and sex; however, the EDMOSFET calculations were discrete in nature, and each phantom only reflected an average body habitus that could not account for the diverse body sizes typical of a pediatric patient population. The dose savings quoted in this study should be considered in light of the CT scan parameters listed in this work and are representative of the average patient habitus emulated by the anthropomorphic phantoms only.

Patient exposure data were calculated from examinations that were performed on 3 different scanners used during the decade (2002-2011) of this retrospective review. Technology rapidly advanced, emphasis on ALARA and dose-reduction techniques in CT became a key focus of international patient safety efforts, and our practices and techniques have adapted and changed over time and continue to do so as low-dose CT technology evolves.

In conclusion, our observations do not support the routine use of pelvic surveillance imaging for patients with Wilms tumor. Omitting pelvic CT from the routine, off-therapy follow-up of patients with Wilms tumor saves an average 30% to 45% of the ED without compromising disease detection. The use of this modality in patients who have high-risk features for pelvic recurrence or pelvic disease, such as age ≥48 months at diagnosis, higher stage disease (stage III), specimen weight >1000 g, positive tumor margins, capsular penetration,14 tumor spillage, and unfavorable histology,25 deserves investigation in prospective national clinical trials.

FUNDING SOURCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

This work was supported in part by grant P30 CA-21765 from the National Institutes of Health, by a Center of Excellence grant from the State of Tennessee, and by the American Lebanese Syrian Associated Charities (ALSAC).

CONFLICT OF INTEREST DISCLOSURES

The authors made no disclosures.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES