Sydney Adventist Hospital uses AI to Increase Cancer Multidisciplinary Meetings Efficiency

Sydney, NSW, (June 12, 2020) – Sydney Adventist Hospital (The San) is now using Artificial Intelligence to automatically populate their cancer Multidisciplinary Team meeting system with pathology information using Natural Language Processing from local company Health Language Analytics (HLA).

The San uses HLA’s CliniSearch application to identify relevant data from a pathology HL7 feed provided by Douglass Hanly Moir (DHM) for the San’s cancer patients. The pathology information is used to automatically pre-populate the correct data fields in preparation for the various cancer Multidisciplinary Team (MDT) meetings run at the San.

Associate Professor Gavin Marx, Director of the Integrated Cancer Centre at the San, sees this as a real advantage for the clinical team and patient care processes. “Pathology is a key component of cancer MDT meetings and there is a significant amount of data generated from each pathology report”, he emphasised. “To have the ability to automatically extract and use this data in a structured format in the context of other available information for clinical decision-making at the meeting saves us a significant amount of time and effort, and enables us to use the data for research purposes”, he explained.

To give a sense of the clinical staff time saved in preparation for an MDT meeting and increased data availability at the meeting at the San, there are sixteen different pathology templates currently included in the extraction service as a joint effort by the San, DHM and HLA. The sixteen reports contain a total of 616 data fields, all of which are identified, extracted from the reports and then coded to SNOMED CT using natural language processing before being fed into the MDT system for presentation at a patient discussion. This is now fully automated, uses no critical staff time and prevents manual transcription errors.

The following table presents the current data extraction volumes:

Tumour StreamTemplateNo. of Fields
BreastBreast Carcinoma61
Breast Carcinoma Post-Neoadjuvant Therapy67
Biomarkers15
LungLung Carcinoma41
UrologyProstate Biopsy23
Radical Prostatectomy54
Prostate TURP18
Bladder Cystectomy32
Renal Pelvis & Ureter34
Renal Cell Nephrectomy40
Renal Cell Partial Nephrectomy33
Testis27
Gynae-OncologyCervical Carcinoma42
Hysterectomy45
Ovaries, Fallopian Tubes, or Peritoneum39
Vulval Carcinoma45
TOTALs16616

Professor Warick Delprado, Director of Histopathology at DHM, highlighted the value of this collaboration, “This is a valuable initiative, allowing us to work in partnership on a project that has mutual benefits for the San and all patients, as well as enhancing the efficiency of our pathologists to continue to deliver high quality service for the Multi-Disciplinary teams”.

The San’s CliniSearch application runs on the Horizon technology platform from HLA and currently has over two million reports indexed for a wide range of clinical concepts. All the pathology reports’ free text sections have been coded to SNOMED CT. By automatically indexing, unpacking and coding the pathology reports from DHM, HLA’s AI is able to fill in the pathology data fields in the newly implemented Multidisciplinary Team meeting system, so that each patient discussion is as fully informed as possible.  With the MDT meetings being moved online due to COVID-19, this ensures that all relevant pathology information is available to all participants remotely, and a high level of patient care is still available.

Professor Jon Patrick, CEO and co-founder of HLA said, “we are pleased to work with the San and DHM by adding AI-based natural language processing to their MDT data workflows. This has reduced admin time and improved accuracy while giving far greater depth to the pathology information available at each MDT meeting”.

ICIMS roll-out to MDT streams continues

Meanwhile, the Sydney Adventist Hospital’s Integrated Cancer Centre had quickly responded to the COVID-19 crisis by moving its MDT meetings onto a remote footing. The Integrated Cancer Information Management System (ICIMS) commenced implementation in June 2019, thanks to a generous donation by the San Foundation. Since then, the San has brought the Lung and Breast tumour streams online, and more recently, the Urology and Gynaecological Oncology modules went live and were launched remotely due to the COVID-19 crisis.

The remote set-up enables cancer case-viewing and treatment-planning on a centralised patient cancer record during the online meetings with Surgeons, Medical Oncologists, Radiation Oncologist, Pathologists, Radiologists, Nurses and Allied Health Services, who may be working from premises all across Sydney.

“The transition to online MDT registrations for our prostate patients has been seamless and very important especially during the COVID crisis. Having the ability to see the surgical and pathological data automatically available in the system has been vital to our MDT discussions” emphasised prominent urologist Professor Henry Woo.

Another immediate benefit of ICIMS is that pathologists and radiologists involved in the MDT now get access to the patient lists and required reviews in real-time. Previously a batch process the day before the meeting often left them needing to work late into the night to compile the required data in time for the meeting. In addition, the data are now automatically populated in ICIMS as soon as a pathology report for a San patient is authorised by the pathologist. These abilities also significantly reduce the workload on the MDT meeting coordinator while increasing data richness and accuracy at the meeting.

The ICIMS solution presents the same functionality to each user whether they are working in a combined conference room or logging in remotely. This means that no user retraining is needed and each patient receives the full benefit of the multidisciplinary team no matter where each team member is located.

iCIMS (Innovative Clinical Information Management Systems)has been recently granted a patent for its LATTICE technology platform by the United States Patent and Trade Mark Office.

See PULSE+IT article here.