Artificial intelligence is increasingly being integrated into oncological diagnostics in Indonesia, where specialists report measurable improvements in the assessment of breast cancer subtypes. During a medical conference marking World Cancer Day in Jakarta, clinicians discussed how AI-assisted pathology may refine the evaluation of Human Epidal Growth Factor Receptor 2 (HER2), a critical biomarker guiding targeted therapy decisions.
Breast cancer remains the most commonly diagnosed malignancy among women globally and represents a major public health concern in Indonesia. According to data from GLOBOCAN 2020, Indonesia recorded approximately 65,000 new cases of breast cancer and more than 22,000 related deaths in that year alone. A proportion of these tumours are characterised by overexpression or amplification of HER2, a protein associated with more aggressive disease behaviour.
AI as a Diagnostic Adjunct in Pathology
At the conference in Jakarta, a specialist in anatomical pathology explained that AI systems can assist in interpreting tumour tissue samples, particularly in determining HER2 status. HER2 testing is performed through laboratory analysis of tumour specimens to establish whether a patient may benefit from anti-HER2 targeted therapies, such as trastuzumab-based regimens.
Evidence referenced at the meeting included data presented at the ASCO Annual Meeting 2025, indicating that AI-supported evaluation increased detection of HER2 “ultra-low” expression by approximately 40 per cent compared with conventional manual assessment. Overall diagnostic accuracy reportedly reached around 92 per cent when AI tools were used alongside pathologists.
In addition to improved sensitivity, consistency between observers also appeared to strengthen. Concordance rates among physicians rose from roughly two-thirds to more than four-fifths in challenging categories such as HER2-low and HER2-ultra-low tumours, which can be difficult to classify using standard visual interpretation alone.
Peer-reviewed research in journals including The Lancet Oncology and Nature Medicine has previously highlighted the potential of digital pathology and machine learning algorithms to reduce inter-observer variability and enhance reproducibility in cancer diagnostics.
Clinical Implications for Targeted Treatment in Indonesia
Accurate HER2 classification has direct therapeutic implications. Patients whose tumours demonstrate HER2 overexpression or specific low-expression profiles may be eligible for targeted anti-HER2 therapies, which act by inhibiting signalling pathways that promote tumour growth. Precision in categorisation is therefore essential to ensure appropriate treatment allocation.
An oncologist specialising in haematology and medical oncology in Indonesia noted that AI systems can accelerate tissue analysis and support more timely clinical decision-making. However, he emphasised that the technology functions as a decision-support tool rather than an autonomous authority. Final judgement remains with the treating physician, who must interpret results within the broader clinical context of each patient.
Technology as Support, Not Replacement
The experience reported in Indonesia reflects a broader global movement towards integrating artificial intelligence into cancer care. While early data suggest meaningful gains in accuracy and efficiency, experts consistently stress that AI should complement — not replace — specialist expertise.
As Indonesia continues to address the substantial burden of breast cancer, the adoption of AI-assisted pathology may contribute to more precise diagnosis and improved therapeutic targeting. Ongoing validation through peer-reviewed research and clinical trials will be essential to confirm long-term benefits and ensure equitable implementation across healthcare settings.
In the evolving landscape of digital medicine, Indonesia’s experience underscores the importance of pairing technological innovation with rigorous clinical oversight to enhance patient outcomes safely and responsibly.