Radiotherapy remains one of the most widely used treatments in modern oncology, relied upon by clinicians across tumour types and disease stages. Alongside its therapeutic benefits, however, radiotherapy can place significant stress on the body’s blood-forming system. For decades, clinicians have closely monitored complete blood count parameters during treatment, looking for signs of anaemia, infection risk, or bleeding tendencies that may compromise patient safety or disrupt therapy.
A new study led by Dr Aniwat Berpan and published in the journal Strahlentherapie und Onkologie offers a more nuanced view of how blood parameters behave during radiotherapy. Titled “Linear regression analysis for complete blood count parameters during radiotherapy”, the research suggests that some long-held assumptions about radiation dose and blood toxicity may warrant reconsideration. Importantly, the findings raise the possibility that blood monitoring during radiotherapy could become more individualised and predictive rather than routine and reactive.
The research was conducted at Samut Sakhon Hospital in Samut Sakhon, Thailand, using a large clinical dataset drawn from routine cancer care.
Why blood monitoring matters during cancer radiotherapy
Abnormal blood counts are common in patients undergoing cancer treatment. Anaemia can reduce oxygen delivery to tissues and has been associated with poorer tumour control and survival outcomes in radiotherapy. Leukopenia and neutropenia increase susceptibility to infection, while thrombocytopenia raises the risk of bleeding complications. Together, these conditions can lead to treatment delays, dose reductions, hospital admissions, and a diminished quality of life.
Due to these risks, many radiotherapy departments perform weekly blood tests during treatment. While this approach prioritises safety, it also places a physical and psychological burden on patients, increases healthcare costs, and consumes clinical resources. Not all patients experience clinically meaningful blood changes, raising an important question: can clinicians predict which patients are most likely to develop significant blood abnormalities during radiotherapy?
The study by Berpan and colleagues addresses this question directly by examining how patient characteristics, treatment factors, and baseline blood values influence blood counts over the course of radiotherapy.
Data-driven insight from real-world radiotherapy
The research analysed data from 496 radiotherapy courses delivered between May 2020 and May 2022. A total of 1,884 complete blood count results were included on a weekly basis. Patients represented a broad range of cancer sites, including breast, pelvis, brain, bone, head and neck, and thoracic tumours. Approximately one-fifth of the courses involved concurrent chemoradiotherapy, a known risk factor for haematological toxicity.
Baseline blood values were measured one week before treatment, followed by weekly testing until radiotherapy was completed. The investigators examined haemoglobin levels, white blood cell count, neutrophils, absolute neutrophil count, and platelet count. Using multiple linear regression models, they assessed how these parameters changed in relation to age, sex, concurrent chemoradiotherapy, treatment site, cumulative radiation dose, and baseline values.
Predictive models for these parameters could guide the need for clinical intervention, facilitate individualized care plans, and reduce unnecessary expenses.
-Aniwat Berpan
Haemoglobin tells a different story than expected
One of the most striking findings of the study relates to haemoglobin, a key indicator of anaemia. The researchers found that haemoglobin levels during radiotherapy were strongly predicted by baseline haemoglobin. For every 1 g per dL increase in baseline haemoglobin, haemoglobin during treatment increased by approximately 0.73 g per dL.
This relationship remained consistent across both development and validation cohorts, with the regression model explaining up to 70 percent of the variation in haemoglobin levels during treatment. In contrast, cumulative radiation dose showed no significant association with haemoglobin changes.
This finding challenges a common assumption in oncology that increasing radiation exposure inevitably worsens anaemia. Instead, the results suggest that haemoglobin levels are relatively stable during radiotherapy and are more dependent on the patient’s starting point than on the amount of radiation delivered.
White cells and platelets follow different biological rules
Unlike haemoglobin, white blood cells, neutrophils, and platelets showed weaker predictive performance in the regression models. Although baseline values were still positively correlated with levels during treatment, the proportion of variance explained was substantially lower.
This difference reflects the biology of blood cell turnover. Red blood cells have a lifespan of approximately 120 days, allowing time for bone marrow outside the radiation field to compensate for reduced production. Neutrophils and platelets, however, have much shorter lifespans, measured in days rather than months. As a result, they are more sensitive to acute bone marrow suppression and systemic stressors such as concurrent chemotherapy.
The study also found negative correlations between cumulative radiation dose and white blood cell and platelet counts, even though these associations were weaker than expected. This suggests that while dose may influence these parameters, it does so in a more complex and less predictable manner.
The role of age, sex, and concurrent chemotherapy
Patient-specific factors also influenced blood count behaviour. Older age was associated with a slight reduction in haemoglobin and platelet levels, although the effect size was small. Male sex was associated with higher haemoglobin levels during treatment, consistent with known physiological differences.
Concurrent chemoradiotherapy was associated with a modest but statistically significant reduction in haemoglobin. This finding aligns with existing evidence that combined modality treatment increases haematological toxicity and reinforces the need for closer monitoring in this patient group.
Notably, the study population included relatively few elderly patients aged 70 years or older. As a result, the findings regarding age should be interpreted cautiously and may not fully reflect risk patterns in older populations.
Challenging established assumptions in radiation oncology
Previous studies, particularly those using intensity-modulated radiotherapy and detailed dose-volume analyses, have linked bone marrow irradiation to anaemia and other blood toxicities. The current study differs in several important ways. All treatments were delivered using three-dimensional conformal radiotherapy, and the cumulative dose was assessed using equivalent dose in 2 Gy fractions rather than detailed marrow dose metrics.
Despite these differences, the findings add to a growing body of evidence suggesting that haemoglobin behaves differently from other blood parameters during radiotherapy. The relative stability of haemoglobin may explain why some patients tolerate treatment without clinically significant anaemia, even when large radiation volumes are involved.
This does not negate the importance of dosimetric planning or bone marrow sparing techniques. Rather, it highlights the need for a tailored risk assessment that considers baseline physiology, treatment site, and concurrent therapies.
Towards predictive and personalised blood monitoring
One of the most compelling implications of this research lies in its potential to reshape clinical practice. If haemoglobin levels during radiotherapy can be reliably predicted using baseline values and simple clinical factors, routine weekly testing may not be necessary for all patients.
Predictive models could help identify patients at low risk of anaemia who may require less frequent monitoring, while focusing attention on those at higher risk who may benefit from early intervention. Such an approach could reduce patient burden, streamline workflows, and lower healthcare costs without compromising safety.
The authors also note that more advanced modelling techniques, including machine learning, may further improve predictive accuracy for white blood cells and platelets. As digital health tools become increasingly integrated into oncology care, data-driven prediction may become a routine part of radiotherapy planning.
Reference
Berpan, A., & Janhom, N. (2025). Linear regression analysis for complete blood count parameters during radiotherapy. Strahlentherapie und Onkologie, 201(5), 561–566. https://doi.org/10.1007/s00066-024-02344-1
