What challenges did Munich Leukemia Lab face before AI implementation?

WHO classification includes many leukaemia and lymphoma sub-types, making diagnosis complex and time-sensitive. Manual transfer of structured questionnaire data into databases and production of patient reports were slow and labor-intensive.

What was the impact of the AI solution?

Specific acute leukaemia subtypes can be diagnosed within about four hours. Many questions can be answered within the first 24 hours, while longest-running methods can take up to a week. Faster and more accurate diagnostic decision-making supports timely treatment and frees expert time for patient care.

Source

MLL Munich Leukemia Laboratory uses AI on AWS to speed leukaemia/lymphoma diagnostics and automate report creation

Use case typeContent generationUpdated Jun 18, 2026

MLL Munich Leukemia Laboratory (Germany) diagnoses blood cancers using AI on AWS and a paperless cloud architecture. The lab handles complex leukaemia and lymphoma classification, large genomic datasets, and patient-specific reporting workflows. It uses large language models and AI-based algorithms with human-in-the-loop oversight to automate report writing and questionnaire ingestion.

Organization: Munich Leukemia Lab
Industry: Healthcare
Location: Germany
Published: June 2026

Reported outcomes

145,000 samples per year

annual samplesOther quantified impact

4 hoursdiagnosis turnaround time24 hoursresponse time for many questions7 dayslongest-running method duration4.5 petabytesgenomic data volume1.4 million casescases investigated

Strategic outcomes

Other strategic outcomeImproved treatment timing for blood cancer patientsOther strategic outcomeAutomated manual questionnaire transfer into the databaseCost efficiencyAvoided the effort of building equivalent high-security infrastructure in-houseScale & capacityEnabled future multi-hub expansion on one cloud system

Primary read

Use case focus

Showing 2 of 2

1Content generation
2Workflow automation

WHO classification includes many leukaemia and lymphoma sub-types, making diagnosis complex and time-sensitive.
Manual transfer of structured questionnaire data into databases and production of patient reports were slow and labor-intensive.

MLL moved to a paperless, cloud-enabled architecture on AWS to handle very large genomic datasets and run AI workloads.
Its AI team uses models to calculate gene variant analyses, generate reports with large language models, and automate questionnaire data ingestion into the database with human-in-the-loop oversight.

Technologies

AWS Large Language Models AI-based algorithms

Specific acute leukaemia subtypes can be diagnosed within about four hours.
Many questions can be answered within the first 24 hours, while longest-running methods can take up to a week.
Faster and more accurate diagnostic decision-making supports timely treatment and frees expert time for patient care.

Sources & evidence1

Groundedness: 4/5Type: Customer StoryPublished: Jun 18, 2026Publisher: AWSEvidence: PrimaryConfidence: High

Primary source

AI-generated summary. Verify important details with the linked sources before relying on this case.

Explore related AI use cases

Customer evidence