Article Text
PDF
Abstract
Malawi is one of the world's poorest countries, but despite this, has a dedicated paediatric oncology service. The service has been hampered by the inability to make a timely cytological diagnosis in the majority of patients. A telemedicine programme was commenced to help overcome this problem, and the results for the first 197 consecutive patients are described. The results are compared with the local reports where available. Most samples were fine needle aspirates (104/197–53%), but others included bone marrow aspirates, peripheral blood films and other fluid collections. A diagnosis was arrived at in 52% of the samples; there were 46 discordant results, 38 were when one or other of the local or distant teams were unable to make a diagnosis, and only 8 where the diagnoses of the 2 teams differed. Diagnoses were made and reports were compiled by the ‘distant’ team within 24 h and sent to the centre in Malawi. This simple telepathology initiative has had a positive impact on clinical management, and could be used in other less resourced centres twinned with better resourced ones.
- Haemato-Oncology
- Histopathology
- Paediatric Haematology
Statistics from Altmetric.com
Introduction
Malawi is one of the world's poorest countries, but has a number of large government hospitals, one of which is the Queen Elizabeth Central Hospital (QECH) in Blantyre. The dedicated paediatric oncology unit sees between 160 and 180 new patients a year. Almost half the children presenting to the unit have Burkitt lymphoma, the rest have the full range of paediatric oncology conditions. Diagnosis is usually made on the basis of a fine needle aspirate (FNA) and/or bone marrow aspirate (BMA). Unfortunately, there is often a long delay in receiving cytology reports as there are very few local pathologists—4 in all, serving a population of 14 million—and they are overwhelmed with the workload of providing a service to the entire country. None of the pathologists are trained specifically in paediatric pathology or haematopathology. Additionally, specialised stains are not available, and the vast majority of diagnoses are made on the basis of Romanowsky stained slides, or in the case of cytological specimens, Giemsa or Papaniculou stains.
As part of the longstanding partnership between the oncology team in Malawi and the team in Newcastle, we have developed a number of initiatives to aid in diagnosis and treatment of children with cancer in Blantyre.1 ,2 After discussion between our units, a relatively inexpensive yet reliable method of rapid diagnosis of samples using telepathology was conceived.
Internet access, a fundamental requirement for telemedicine has grown rapidly in Africa, including Malawi, over the past few decades, and now all 54 African countries have direct internet access in the major cities.3 There are many ways of using telepathology, ranging from complicated, remotely driven microscopes4 to simple e-mails of photographic images. It is important that the provision of telepathology services is driven by the need of the local centre, that appropriate clinical information is provided and that the results are interpreted for clinical use by the local clinician who understands the limitation of the diagnosis that such a system imposes. Such a system needs to be robust and deliverable; and the system needs to be established in light of the units’ infrastructure and needs.
The methods, outcomes and impact of the initial year of this initiative are described below.
Methods
Using an existing Olympus microscope (CX31), an additional viewing head was purchased along with a specialist Jenoptik microscope camera (CT13) with accompanying software (Prog Res Mac Capture pro). The camera was attached to one of the viewing heads of the microscope and connected to an Apple Macintosh Macbook pro (2011) via a firewire connector. The slides were prepared by spreading and air-drying at the patient's bedside, and then stained using a rapid stain (Rapidiff stain kit by Clinical Services Laboratories). The microscope operator (usually EM) chose the portions of each slide to photograph and added the clinical details (range 3–6 photographs: average 4). The images were set up to be automatically saved into a shared Dropbox with the clinical details attached and automatic notification to the receiving team. The images were looked at by the distant team within 24 h, and a report e-mailed to the local clinician, as well as saved to the Dropbox (figure 1). There were no problems with the uploading or viewing of files, and the photographs and reports were saved hourly by Apple time-capsule technology.
Example of a report sent to the team in Malawi.
Patients
The first 197 consecutive cytology samples reviewed in the system from the paediatric oncology ward at the QECH are described. These were seen over a 1-year period from July 2012 to 2013. The age range of the patients was 1 month—15 years (7 unknown). There were 96 females and 101 males. The sample types and rate of diagnosis are tabulated below (table 1).
The number and types of samples analysed and the percentage where a diagnosis was made
Results
Of the 197 samples, there was no local histology report for 70 patients. Where local diagnoses were made, the time from sample collection to report ranged from 3 weeks to 4 months. Reasons for the delay are mainly due to the excessive workload of the few pathologists in Malawi rather than any inherent problems in the processing of the samples.
Most of the samples were FNAs, in keeping with usual local diagnostic practice. These accounted for 53% (104) of the samples examined, and resulted in a diagnosis being made in 48% of those examined. A further 25 (13%) samples had FNA and BMA available, and diagnoses were made in 75% (20) of cases. A diagnosis was able to be made in 84% (16/19) of the BMAs. The rest of the sample types had more disappointing results and are summarised in table 1.
Of the 127 samples in which local and distant pathology were available, 81 were concordant in their diagnosis, in 38 a diagnosis was made by one centre (and none in the other) and 8 had conflicting diagnoses (table 2). Of the 81 which were concordant, 35 had Burkitt lymphoma, in 32 no diagnosis was possible (primarily due to poor cellularity or blood contamination of FNA), 5 had a small round blue cell tumour of varying aetiologies, 4 had acute lymphoblastic leukaemia (ALL), 2 had Wilms tumour and 1 each had non-Hodgkin's lymphoma (NHL), neuroblastoma and chronic myeloid leukaemia. Of the 38 where a diagnosis was made by one centre and no diagnosis was suggested by the other, 8 had a diagnosis of Burkitt Lymphoma, 6 a sarcoma and 5 a diagnosis of malignancy, not otherwise specified, which was made locally: 5 had a diagnosis of ALL made distantly. Of the remaining 14 samples, 8 had a local diagnosis but not a distant one, and 6 vice versa. Where there was a discordant diagnosis, the treatment was determined clinically or on the basis of the first available cytology. These cases, the treatment received and their outcomes are outlined in table 2.
Non-concordant samples—details
Discussion
Treating children with cancer in resource-limited countries is challenging due to the lack of detailed investigations, appropriate chemotherapy agents and an inadequate supportive care infrastructure, but also because of the impaired ability to make a robust histological diagnosis in a timely manner. These problems faced by the team in Malawi resulted in a simple telepathology initiative that has enabled a rapid histological diagnostic review for the children with potential cancer in Blantyre, Malawi, and enabled appropriate selection of treatment.
It is important when choosing a telepathology initiative, that it is appropriate for the environment. Although other methods, such as video and use of YouTube to upload these, may have been possible, we decided to keep the system as simple as possible to reduce potential problems. Other published studies have used a variety of methods including Skype,5 web-based systems, such as Campus Medicus and ipath.6 In Malawi, where power outages are common, the internet speed is slow, and it is problematic to service and maintain complex machinery, a system was chosen that would be relatively immune to these problems, but maintained efficacy and simplicity appropriate to the skills the team in Malawi and Newcastle possessed. Using a rapid and simple Rapidiff staining system ensured good quality slides, the use of an easy-to-use, yet good quality microscope and camera, yielded adequate photographs. The transmission system was simple and not prone to failure. A more complex scanning system would have resulted in larger files, the need for complex servicing, and no one on the spot to fix the machine if it developed any technical issues. The first 50 slides were reviewed retrospectively after transport to Newcastle, and there was only one where it would have been possible to make a diagnosis by seeing the whole slide instead of the selected photographs. There is a degree of learned skill needed to photograph appropriate microscopic fields from the slides, but this was achieved by the referring clinicians with a high degree of competence.
The most important aspect in applying the results is for the local clinicians to interpret them in their clinical context. Additionally, complex cases were discussed either via e-mail or teleconference (Africa-shared care meeting via Cure 4 Kids teleconferencing facility). Occasionally, slides were sent to Newcastle for more detailed testing. There is approximately a 50% diagnostic rate which at first glance does not seem ideal. However, even in the best hands, FNAs do not have an optimal diagnostic yield (approximately 80% if aspirates are cellular). Added to this, the inability to perform more complex stains, and to view the slides as a whole, reduces the ability to make a diagnosis.
Despite these limitations, the discordance rate was low with only eight samples resulting in potentially incorrect treatment. This system has resulted in a dramatic increase in the number of children able to have treatment based on a cytological diagnosis, as delays in local pathological diagnosis inevitably meant that treatment had to be initiated prior to this being available.
The telemedicine initiative described above has resulted in an improved standard of care for children with cancer in Blantyre. This model could easily be reproduced in other settings where there is a twinning arrangement between less well-resourced and well-resourced centres. In order to improve the diagnostic ability in Malawi, formal biopsy and a resulting complexity in tissue preparation, preservation and staining will be necessary.
Take home messages
-
Simple telemedicine especially when there are delays in local diagnosis are useful to the treating clinical team.
-
Telemedicine initiatives do not need to be complicated.
-
Simple and rapid telemedicine provides an accurate diagnosis in the vast majority of cases.
Footnotes
-
CP and FR contributed equally.
MEM and BS contributed equally
-
Correction notice V Wadehra's name has been corrected and a sentence under 'Patients' has been modified since this article was published Online First.
-
Contributors Conception of study: SB, EMM, PC. Collection and processing samples: RF, EMM, GC, KB. Examination of material: PC, VW, SS, GL, TT. Collection of clinical data: DS, EMM, SB. Writing of manuscript: SB, EMM, PC, RF. Review of manuscript: all.
-
Competing interests None.
-
Provenance and peer review Not commissioned; externally peer reviewed.
-
Data sharing statement This is not appropriate for our study as these images are not research samples but merely transmitted images which are the property of the referring hospital in Malawi.