MRI Characteristics of Pediatric Renal Tumors: A SIOP-RTSG Radiology Panel Delphi Study.

BACKGROUND: The SIOP-Renal Tumor Study Group (RTSG) does not advocate invasive procedures to determine histology before the start of therapy. This may induce misdiagnosis-based treatment initiation, but only for a relatively small percentage of approximately 10% of non-Wilms tumors (non-WTs). MRI could be useful for reducing misdiagnosis, but there is no global consensus on differentiating characteristics. PURPOSE: To identify MRI characteristics that may be used for discrimination of newly diagnosed pediatric renal tumors. STUDY TYPE: Consensus process using a Delphi method. POPULATION: Not applicable. FIELD STRENGTH/SEQUENCE: Abdominal MRI including T1- and T2-weighted imaging, contrast-enhanced MRI, and diffusion-weighted imaging at 1.5 or 3 T. ASSESSMENT: Twenty-three radiologists from the SIOP-RTSG radiology panel with ≥5 years of experience in MRI of pediatric renal tumors and/or who had assessed ≥50 MRI scans of pediatric renal tumors in the past 5 years identified potentially discriminatory characteristics in the first questionnaire. These characteristics were scored in the subsequent second round, consisting of 5-point Likert scales, ranking- and multiple choice questions. STATISTICAL TESTS: The cut-off value for consensus and agreement among the majority was ≥75% and ≥60%, respectively, with a median of ≥4 on the Likert scale. RESULTS: Consensus on specific characteristics mainly concerned the discrimination between WTs and non-WTs, and WTs and nephrogenic rest(s) (NR)/nephroblastomatosis. The presence of bilateral lesions (75.0%) and NR/nephroblastomatosis (65.0%) were MRI characteristics indicated as specific for the diagnosis of a WT, and 91.3% of the participants agreed that MRI is useful to distinguish NR/nephroblastomatosis from WT. Furthermore, all participants agreed that age influenced their prediction in the discrimination of pediatric renal tumors. DATA CONCLUSION: Although the discrimination of pediatric renal tumors based on MRI remains challenging, this study identified some specific characteristics for tumor subtypes, based on the shared opinion of experts. These results may guide future validation studies and innovative efforts. LEVEL OF EVIDENCE: 3 Technical Efficacy Stage: 3.

Relapse of Wilms' tumour and detection methods: a retrospective analysis of the 2001 Renal Tumour Study Group-International Society of Paediatric Oncology Wilms' tumour protocol database.

BACKGROUND: Wilms' tumour is the most common renal cancer in childhood and about 15% of patients will relapse. There is scarce evidence about optimal surveillance schedules and methods for detection of tumour relapse after therapy. METHODS: The Renal Tumour Study Group-International Society of Paediatric Oncology (RTSG-SIOP) Wilms' tumour 2001 trial and study is an international, multicentre, prospective registration, biological study with an embedded randomised clinical trial for children with renal tumours aged between 6 months and 18 years. The study covers 243 different centres in 27 countries grouped into five consortia. The current protocol of SIOP surveillance for Wilms' tumour recommends that abdominal ultrasound and chest x-ray should be done every 3 months for the first 2 years after treatment and be repeated every 4-6 months in the third and fourth year and annually in the fifth year. In this retrospective cohort study of the protocol database, we analysed data from participating institutions on timing, anatomical site, and mode of detection of all first relapses of Wilms' tumour. The primary outcomes were how relapse of Wilms' tumour was detected (ie, at or between scheduled surveillance and with or without clinical symptoms, scan modality, and physical examination) and to estimate the number of scans needed to capture one subclinical relapse. The RTSG-SIOP study is registered with Eudra-CT, number 2007-004591-39. FINDINGS: Between June 26, 2001, and May 8, 2015, of 4271 eligible patients in the 2001 RTSG-SIOP Wilms' tumour database, 538 (13%) relapsed. Median follow-up from surgery was 62 months (IQR 32-93). The method used to detect relapse was registered for 410 (76%) of 538 relapses. Planned surveillance imaging captured 289 (70%) of these 410 relapses. The primary imaging modality used to detect relapse was reported for 251 patients, among which relapse was identified by abdominal ultrasound (80 [32%] patients), chest x-ray (78 [31%]), CT scan of the chest (64 [25%]) or abdomen (20 [8%]), and abdominal MRI (nine [4%]). 279 (68%) of 410 relapses were not detectable by physical examination and 261 (64%) patients did not have clinical symptoms at relapse. The estimated number of scans needed to detect one subclinical relapse during the first 2 years after nephrectomy was 112 (95% CI 106-119) and, for 2-5 years after nephrectomy, 500 (416-588). INTERPRETATION: Planned surveillance imaging captured more than two-thirds of predominantly asymptomatic relapses of Wilms' tumours, with most detected by abdominal ultrasound, chest x-ray, or chest CT scan. Beyond 2 years post-nephrectomy, a substantial number of surveillance scans are needed to capture one relapse, which places a burden on families and health-care systems. FUNDING: Great Ormond Street Hospital Children's Charity, the European Expert Paediatric Oncology Reference Network for Diagnostics and Treatment, The Danish Childhood Cancer Foundation, Cancer Research UK, the UK National Cancer Research Network and Children's Cancer and Leukaemia Group, Société Française des Cancers de l'Enfant and Association Leon Berard Enfant Cancéreux and Enfant et Santé, Gesellschaft für Pädiatrische Onkologie und Hämatologie and Deutsche Krebshilfe, Grupo Cooperativo Brasileiro para o Tratamento do Tumor de Wilms and Sociedade Brasileira de Oncologia Pediátrica, the Spanish Society of Pediatric Haematology and Oncology and the Spanish Association Against Cancer, and SIOP-Netherlands.

Detecting inflammation in inflammatory bowel disease - how does ultrasound compare to magnetic resonance enterography using standardised scoring systems?

BACKGROUND: Magnetic resonance enterography (MRE) is the current gold standard for imaging in inflammatory bowel disease, but ultrasound (US) is a potential alternative. OBJECTIVE: To determine whether US is as good as MRE for the detecting inflamed bowel, using a combined consensus score as the reference standard. MATERIALS AND METHODS: We conducted a retrospective cohort study in children and adolescents <18 years with inflammatory bowel disease (IBD) at a tertiary and quaternary centre. We included children who underwent MRE and US within 4 weeks. We scored MRE using the London score and US using a score adapted from the METRIC (MR Enterography or Ultrasound in Crohn's Disease) trial. Four gastroenterologists assessed an independent clinical consensus score. A combined consensus score using the imaging and clinical scores was agreed upon and used as the reference standard to compare MRE with US. RESULTS: We included 53 children. At a whole-patient level, MRE scores were 2% higher than US scores. We used Lin coefficient to assess inter-observer variability. The repeatability of MRE scores was poor (Lin 0.6). Agreement for US scoring was substantial (Lin 0.95). There was a significant positive correlation between MRE and clinical consensus scores (Spearman's rho = 0.598, P=0.0053) and US and clinical consensus scores (Spearman's rho = 0.657, P=0.0016). CONCLUSION: US detects as much clinically significant bowel disease as MRE. It is possible that MRE overestimates the presence of disease when using a scoring system. This study demonstrates the feasibility of using a clinical consensus reference standard in paediatric IBD imaging studies.

Whole-tumor apparent diffusion coefficient measurements in nephroblastoma: Can it identify blastemal predominance?

PURPOSE: To explore the potential relation between whole-tumor apparent diffusion coefficient (ADC) parameters in viable parts of tumor and histopathological findings in nephroblastoma. MATERIALS AND METHODS: Children (n = 52) with histopathologically proven nephroblastoma underwent diffusion-weighted magnetic resonance imaging (MRI) (1.5T) before preoperative chemotherapy. Of these, 25 underwent an additional MRI after preoperative chemotherapy, shortly before resection. An experienced reader performed the whole-tumor ADC measurements of all lesions, excluding nonenhancing areas. An experienced pathologist reviewed the postoperative specimens according to standard SIOP guidelines. Potential associations between ADC parameters and proportions of histological subtypes were assessed with Pearson's or Spearman's rank correlation coefficient depending on whether the parameters tested were normally distributed. In case the Mann-Whitney U-test revealed significantly different ADC values in a subtype tumor, this ADC parameter was used to derive a receiver operating characteristic (ROC) curve. RESULTS: The 25th percentile ADC at presentation was the best ADC metric correlated with proportion of blastema (Pearson's r = -0.303, P = 0.026). ADC after preoperative treatment showed moderate correlation with proportion stromal subtype at histopathology (r = 0.579, P = 0.002). By ROC analysis, the optimal threshold of median ADC for detecting stromal subtype was 1.362 × 10-3 mm2 /s with sensitivity and specificity of 100% (95% confidence interval [CI] 0.65-1.00) and 78.9% (95% CI 0.57-0.92), respectively. CONCLUSION: ADC markers in nephroblastoma are related to stromal subtype histopathology; however, identification of blastemal predominant tumors using whole-tumor ADC measurements is probably not feasible. LEVEL OF EVIDENCE: 3 J. MAGN. RESON. IMAGING 2017;45:1316-1324.

Paediatric magnetic resonance enteroclysis under general anaesthesia - initial experience.

MR enterography is the accepted imaging reference standard for small bowel assessment in inflammatory bowel disease. There is an increasing cohort of children with inflammatory bowel disease presenting at an early age (<5 years) with severe disease. Younger children present a technical challenge for enterography because of the need for sedation/general anaesthesia to allow image optimisation and the need for oral contrast to allow adequate luminal assessment. Through our experiences, MR enteroclysis under general anaesthesia has proven to be a successful imaging technique for the work-up of these patients. In this paper, we present our institutional practice for performing MR enteroclysis under general anaesthesia.

Apparent diffusion coefficient as it relates to histopathology findings in post-chemotherapy nephroblastoma: a feasibility study.

BACKGROUND: Nephroblastomas represent a group of heterogeneous tumours with variable proportions of distinct histopathological components. OBJECTIVE: The purpose of this study was to investigate whether direct comparison of apparent diffusion coefficient (ADC) measurements with post-resection histopathology subtypes is feasible and whether ADC metrics are related to histopathological components. MATERIALS AND METHODS: Twenty-three children were eligible for inclusion in this retrospective study. All children had MRI including diffusion-weighted imaging (DWI) after preoperative chemotherapy, just before tumour resection. A pathologist and radiologist identified corresponding slices at MRI and postoperative specimens using tumour morphology, the upper/lower calyx and hilar vessels as reference points. An experienced reader performed ADC measurements, excluding non-enhancing areas. A pathologist reviewed the corresponding postoperative slides according to the international standard guidelines. We tested potential associations with the Spearman rank test. RESULTS: Side-by-side comparison of MRI-DWI with corresponding histopathology slides was feasible in 15 transverse slices in 9 lesions in 8 patients. Most exclusions were related to extensive areas of necrosis/haemorrhage. In one lesion correlation was not possible because of the different orientation of sectioning of the specimen and MRI slices. The 25% ADC showed a strong relationship with percentage of blastema (Spearman rho=-0.71, P=0.003), whereas median ADC was strongly related to the percentage stroma (Spearman rho=0.74, P=0.002) at histopathology. CONCLUSION: Side-by-side comparison of MRI-DWI and histopathology is feasible in the majority of patients who do not have massive necrosis and hemorrhage. Blastemal and stromal components have a strong linear relationship with ADC markers.

Comparison of diagnostic performance for perinatal and paediatric post-mortem imaging: CT versus MRI.

OBJECTIVES: To compare the diagnostic yield of whole-body post-mortem computed tomography (PMCT) imaging to post-mortem magnetic resonance (PMMR) imaging in a prospective study of fetuses and children. METHODS: We compared PMCT and PMMR to conventional autopsy as the gold standard for the detection of (a) major pathological abnormalities related to the cause of death and (b) all diagnostic findings in five different body organ systems. RESULTS: Eighty two cases (53 fetuses and 29 children) underwent PMCT and PMMR prior to autopsy, at which 55 major abnormalities were identified. Significantly more PMCT than PMMR examinations were non-diagnostic (18/82 vs. 4/82; 21.9 % vs. 4.9 %, diff 17.1 % (95 % CI 6.7, 27.6; p < 0.05)). PMMR gave an accurate diagnosis in 24/55 (43.64 %; 95 % CI 31.37, 56.73 %) compared to 18/55 PMCT (32.73 %; 95 % CI 21.81, 45.90). PMCT was particularly poor in fetuses <24 weeks, with 28.6 % (8.1, 46.4 %) more non-diagnostic scans. Where both PMCT and PMMR were diagnostic, PMMR gave slightly higher diagnostic accuracy than PMCT (62.8 % vs. 59.4 %). CONCLUSION: Unenhanced PMCT has limited value in detection of major pathology primarily because of poor-quality, non-diagnostic fetal images. On this basis, PMMR should be the modality of choice for non-invasive PM imaging in fetuses and children. KEY POINTS: • Overall 17.1 % more PMCT examinations than PMMR were non-diagnostic • 28.6 % more PMCT were non-diagnostic than PMMR in fetuses <24 weeks • PMMR detected almost a third more pathological abnormalities than PMCT • PMMR gave slightly higher diagnostic accuracy when both were diagnostic.

A multi-Gaussian model for apparent diffusion coefficient histogram analysis of Wilms' tumour subtype and response to chemotherapy.

Wilms' tumours (WTs) are large heterogeneous tumours, which typically consist of a mixture of histological cell types, together with regions of chemotherapy-induced regressive change and necrosis. The predominant cell type in a WT is assessed histologically following nephrectomy, and used to assess the tumour subtype and potential risk. The purpose of this study was to develop a mathematical model to identify subregions within WTs with distinct cellular environments in vivo, determined using apparent diffusion coefficient (ADC) values from diffusion-weighted imaging (DWI). We recorded the WT subtype from the histopathology of 32 tumours resected in patients who received DWI prior to surgery after pre-operative chemotherapy had been administered. In 23 of these tumours, DWI data were also available prior to chemotherapy. Histograms of ADC values were analysed using a multi-Gaussian model fitting procedure, which identified 'subpopulations' with distinct cellular environments within the tumour volume. The mean and lower quartile ADC values of the predominant viable tissue subpopulation (ADC(1MEAN), ADC(1LQ)), together with the same parameters from the entire tumour volume (ADC(0MEAN), ADC(0LQ)), were tested as predictors of WT subtype. ADC(1LQ) from the multi-Gaussian model was the most effective parameter for the stratification of WT subtype, with significantly lower values observed in high-risk blastemal-type WTs compared with intermediate-risk stromal, regressive and mixed-type WTs (p < 0.05). No significant difference in ADC(1LQ) was found between blastemal-type and intermediate-risk epithelial-type WTs. The predominant viable tissue subpopulation in every stromal-type WT underwent a positive shift in ADC(1MEAN) after chemotherapy. Our results suggest that our multi-Gaussian model is a useful tool for differentiating distinct cellular regions within WTs, which helps to identify the predominant histological cell type in the tumour in vivo. This shows potential for improving the risk-based stratification of patients at an early stage, and for guiding biopsies to target the most malignant part of the tumour.

Why measure tumours?

This article questions the scientific justification of ingrained radiologic practices exemplified by size measurements of childhood solid tumours. This is approached by a critical review of staging systems from a selection of paediatric oncological treatment protocols. Local staging remains size-dependent for some tumour types. The consequent stage assignment can significantly influence treatment intensity. Still, the protocols tend not to give precise guidance on how to perform scans and standardise measurements. Also, they do not estimate or account for the inevitable variability in measurements. Counts and measurements of lung nodules are, within some tumour groups, used for diagnosis of metastatic disease. There is, however, no evidence that nodule size is a useful discriminator of benign and malignant lung nodules. The efficacy of imaging depends chiefly on observations being precise, accurate and valid for the desired diagnostic purpose. Because measurements without estimates of their errors are meaningless, studies of variability dependent on tumour shape and location, imaging device and observer need to be encouraged. Reproducible observations make good candidates for staging parameters if they have prognostic validity and at the same time show little covariation with (thereby adding new information to) the existing staging system. The lack of scientific rigour has made the validity of size measurement very difficult to assess. Action is needed, the most important being radiologists' active contribution in development of oncological staging systems, attention to standardisation, knowledge about errors in measurement and protection against undue influence of such errors in the staging of the individual child.

Fusion and subtraction post-processing in body MRI.

Interpreting complex paediatric body MRI studies requires the integration of information from multiple sequences. Image processing software, some freely available, allows the radiologist to use simple and rapid post-processing techniques that may aid diagnosis. We demonstrate the use of fusion and subtraction post-processing techniques with examples from four areas of application: enterography, oncological imaging, musculoskeletal imaging and MR fistulography.

Intra- and interobserver variability of whole-tumour apparent diffusion coefficient measurements in nephroblastoma: a pilot study.

BACKGROUND: The apparent diffusion coefficient (ADC) is potentially useful for assessing treatment response in nephroblastoma (Wilms tumour). However the precision of ADC measurements in these heterogeneous lesions is unknown. OBJECTIVE: To assess intra- and interobserver variability of whole-tumour ADC measurements in viable parts of nephroblastomas at diagnosis and after preoperative chemotherapy. MATERIALS AND METHODS: We included children with histopathologically proven nephroblastoma who had undergone MRI with diffusion-weighted imaging before and after preoperative chemotherapy. Three independent observers performed whole-tumour ADC measurements of all lesions, excluding non-enhancing areas. One observer evaluated all lesions on two occasions. We performed analyses using Bland-Altman plots and concordance correlation coefficient (CCC) calculations with 95% limits of agreement for median ADC, difference between pre- and post-chemotherapy median ADC (ADC shift) and percentage of pixels with ADC values <1.0 × 10(-3) mm(2)/s. RESULTS: In 22 lesions (13 pretreatment and 9 post-treatment) in 10 children the interobserver variability in median ADC and ADC shift were within the interval of approximately ±0.1 × 10(-3) mm(2)/s (limits of agreement for median ADC ranged -0.08-0.11 × 10(-3) mm(2)/s and for ADC-shift -0.11-0.09 × 10(-3) mm(2)/s). The interobserver variability for percentage of low-ADC pixels was larger and also biased. The calculated CCC confirmed good intra- and interobserver agreement (ρ-c ranging from 0.968 to 0.996). CONCLUSION: Measurements of whole-tumour ADC values excluding necrotic areas seem to be sufficiently precise for detection of chemotherapy-related change.

Diagnostic accuracy of post-mortem MRI for thoracic abnormalities in fetuses and children.

OBJECTIVES: To compare the diagnostic accuracy of post-mortem magnetic resonance imaging (PMMR) specifically for non-cardiac thoracic pathology in fetuses and children, compared with conventional autopsy. METHODS: Institutional ethics approval and parental consent was obtained. A total of 400 unselected fetuses and children underwent PMMR before conventional autopsy, reported blinded to the other dataset. RESULTS: Of 400 non-cardiac thoracic abnormalities, 113 (28 %) were found at autopsy. Overall sensitivity and specificity (95 % confidence interval) of PMMR for any thoracic pathology was poor at 39.6 % (31.0, 48.9) and 85.5 % (80.7, 89.2) respectively, with positive predictive value (PPV) 53.7 % (42.9, 64.0) and negative predictive value (NPV) 77.0 % (71.8, 81.4). Overall agreement was 71.8 % (67.1, 76.2). PMMR was most sensitive at detecting anatomical abnormalities, including pleural effusions and lung or thoracic hypoplasia, but particularly poor at detecting infection. CONCLUSIONS: PMMR currently has relatively poor diagnostic detection rates for the commonest intra-thoracic pathologies identified at autopsy in fetuses and children, including respiratory tract infection and diffuse alveolar haemorrhage. The reasonable NPV suggests that normal thoracic appearances at PMMR exclude the majority of important thoracic lesions at autopsy, and so could be useful in the context of minimally invasive autopsy for detecting non-cardiac thoracic abnormalities. KEY POINTS: • PMMR has relatively poor diagnostic detection rates for common intrathoracic pathology • The moderate NPV suggests that normal PMMR appearances exclude most important abnormalities • Lung sampling at autopsy remains the "gold standard" for pulmonary pathology.

Gadolinium and nephrogenic systemic fibrosis: an update.

Nephrogenic systemic fibrosis (NSF) is a multisystem disease seen exclusively in patients with renal impairment. It can be severely debilitating and sometimes fatal. There is a strong association with gadolinium-based contrast agents used in magnetic resonance imaging (MRI). Risk factors include renal impairment and proinflammatory conditions, e.g. major surgery and vascular events. Although there is no single effective treatment for NSF, the most successful outcomes are seen following restoration of renal function, either following recovery from acute kidney injury or following renal transplantation. There have been ten biopsy-proved pediatric cases of NSF, with no convincing evidence that children have a significantly altered risk compared with the adult population. After implementation of guidelines restricting the use of gadolinium-based contrast agents in at-risk patients, there has been a sharp reduction in new cases and no new reports in children. Continued vigilance is recommended: screening for renal impairment, use of more stable gadolinium chelates, consideration of non-contrast-enhanced MRI or alternative imaging modalities where appropriate.

The accountability for reasonableness approach to guide priority setting in health systems within limited resources--findings from action research at district level in Kenya, Tanzania, and Zambia.

BACKGROUND: Priority-setting decisions are based on an important, but not sufficient set of values and thus lead to disagreement on priorities. Accountability for Reasonableness (AFR) is an ethics-based approach to a legitimate and fair priority-setting process that builds upon four conditions: relevance, publicity, appeals, and enforcement, which facilitate agreement on priority-setting decisions and gain support for their implementation. This paper focuses on the assessment of AFR within the project REsponse to ACcountable priority setting for Trust in health systems (REACT). METHODS: This intervention study applied an action research methodology to assess implementation of AFR in one district in Kenya, Tanzania, and Zambia, respectively. The assessments focused on selected disease, program, and managerial areas. An implementing action research team of core health team members and supporting researchers was formed to implement, and continually assess and improve the application of the four conditions. Researchers evaluated the intervention using qualitative and quantitative data collection and analysis methods. RESULTS: The values underlying the AFR approach were in all three districts well-aligned with general values expressed by both service providers and community representatives. There was some variation in the interpretations and actual use of the AFR in the decision-making processes in the three districts, and its effect ranged from an increase in awareness of the importance of fairness to a broadened engagement of health team members and other stakeholders in priority setting and other decision-making processes. CONCLUSIONS: District stakeholders were able to take greater charge of closing the gap between nationally set planning and the local realities and demands of the served communities within the limited resources at hand. This study thus indicates that the operationalization of the four broadly defined and linked conditions is both possible and seems to be responding to an actual demand. This provides arguments for the continued application and further assessment of the potential of AFR in supporting priority-setting and other decision-making processes in health systems to achieve better agreed and more sustainable health improvements linked to a mutual democratic learning with potential wider implications.

MRI: how to perform a pediatric scan.

Magnetic resonance imaging (MRI) is rich in diagnostic information but requires optimization for use in children. The main problems are motion artifacts and poor signal-to-noise ratio (SNR). SNR is proportional to voxel volume, which must therefore not be too small, however, usually needs to be reduced compared to adult imaging to account for the finer anatomy of the child. The use of multi-channel coils with element sizes appropriate for the anatomy of interest ensures optimal baseline SNR. Longer acquisition time increases SNR (with a square-root factor), but the flip-side is that this allows more motion artifacts. Attention to patient preparation and to techniques for motion artifact reduction is therefore crucial, and the most important principles are discussed. Low SNR may in part be compensated by optimizing the image contrast by weighting (tissue and lesions T1 and T2 may differ from adults) and by using contrast agents. It is also powerful to combine different image contrasts during postprocessing. The basic principles are discussed, followed by an example scan protocol.

Advances in pediatric oncology MRI.

Refined stratification of disease is thought to result in better survival from childhood malignant disease while minimizing the adverse effects of anticancer therapies. There is a potential for magnetic resonance imaging (MRI) to contribute to such stratification by improved tissue characterization, anatomical depiction, staging, and assessment of early treatment response. Recent advances in pediatric MRI outside the central nervous system (CNS) are reviewed in this context. The focus is on new applications for conventional MRI and on clinical implementation of tissue-specific and quantitative techniques. This area is largely unexplored, and potential directions for research are indicated.

Immunisation coverage and its associations in rural Tanzanian infants.

INTRODUCTION: In Tanzania, vaccination rates (VRs) range from 80% to 90% for standard vaccines, but little information is available about rural populations and nomadic pastoralists. This study investigates levels and trends of the immunisation status of infants at eight mobile reproductive-and-child-health (RCH) clinics in a rural area in northern Tanzania (with a large multi-tribal population that has a significant population of nomadic pastoralists) for the years 1998, 1999, 2006 and 2007. In addition, the influence of tribal affiliation and health system-related factors on the immunisation status in this population is analysed. METHODS: Vaccination data of 3868 infants for the standard bacillus Calmette-Guérin (BCG), poliomyelitis, diphtheria, pertussis, tetanus and measles vaccines were obtained from the RCH clinic records retrospectively, and coverage for both single vaccines and full vaccination by the end of first year of life were calculated. These results were correlated with data on predominant tribal affiliation at the clinic site, skilled attendance at birth, service provision and vaccine availability as independent variables. RESULTS: In 1998, the full vaccination rate (FVR) across all RCH clinics was 72%, significantly higher than in the other years (1999: 58%; 2006: 58%; 2007: 57%) (p<0.0001). BCG and measles VRs were highest in 1998 and 1999, whereas VR was lowest for poliomyelitis in 1999, and for diphtheria-pertussis-tetanus in 2007 (all p<0.001). Measles VR showed a declining trend (1998: 72%; 1999: 73%; 2006: 62%; 2007: 59%) affecting the FVR, except in 1999 when poliomyelitis VR was lower (67%). FVR > 80% was only achieved at one clinic during 3 years. No clinic showed a consistent increase of VRs over time. In univariate analysis, predominant tribal affiliation (Datoga tribe) was associated with a low FVR (odds ratio (OR) 4.6 (95% confidence interval (CI) 3.8-5.5)), as were low rates of skilled attendance at birth (OR 3.6 (CI 2.9-4.4)). Other health system-related factors associated with low FVRs included interruption of scheduled monthly immunisation clinics (OR 9.8 (CI 2.1-45.5)) and lack of vaccines (OR 1.2-2.9, depending on vaccine). In multivariate analysis, predominant Datoga tribal affiliation and lack of vaccines retained their association with the risk of low rates of vaccination. CONCLUSIONS: Vaccination rates in this difficult-to-reach population are markedly lower than the national average for almost all years and clinics. Affiliation to the nomadic Datoga tribe and lack of vaccines determine VRs in this rural population. Improvements in immunisation service delivery, vaccine availability, stronger involvement of the nomadic communities and special outreach services for this population are required to improve VRs in these remote areas of Tanzania.