JSJan Seppälä

Jan Seppälä

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Jan Seppälä has a total of 126 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Variability in Breast Radiotherapy Treatment Plan Evaluation: A Multi-Institutional Study of Expert Agreement and Decision-Making  

    by Hiran Maladenige, Janne Heikkilä, Jan Seppälä, Tuomas Virén, Juuso T. J. Honkanen, Akseli Leino, Karoliina Hundlinger, Tero Vatanen, Laura Tuomikoski, Tiina Seppälä, et al.

    Objective: To quantify inter- and intra-rater variability in comparative breast radiotherapy treatment plan evaluation and determine how individual evaluation criteria contribute to overall plan selection. Approach: Nine medical physicists from four radiotherapy centres independently evaluated paired treatment plans (original and systematically modified plans) for 21 left-sided breast cancer patients treated with tangential volumetric modulated arc therapy. Plans were evaluated using target- and organ-at-risk-related criteria in two blinded rounds. Inter- and intra-rater agreement were assessed using Fleiss' and Cohen's kappa, respectively. A mixed-effects logistic regression model was used to quantify the influence of individual evaluation criteria on the final verdict. Main results: Inter-rater agreement varied across evaluation criteria. Organ-specific dose assessments showed high agreement (κ=0.74-0.94), whereas integrative criteria, including target dose homogeneity, overall organ-at-risk dose, control of low-dose spread, and the final verdict, showed lower agreement (κ=0.45-0.63). In contrast, intra-rater agreement remained high (mean κ=0.82), indicating consistent individual evaluation behaviour. Mixed-effects modelling identified a hierarchy of decision-making priorities. Overall organ-at-risk dose, target coverage, and ipsilateral lung dose had the greatest influence on final verdict (odds ratio: 10–16), whereas other criteria contributed less. Criteria with greater influence on decision-making generally showed higher inter-rater agreement. Significance: Radiotherapy plan evaluation is both variable and structured. Variability arises from differences in how evaluators prioritize competing dose-volume factors rather than from inconsistency in individual assessments. The relationship between criterion importance and observer agreement provides insight into clinical decision-making and establishes a reference framework for developing and validating objective treatment plan evaluation methods.

  • Open Access

    KaSARi: a national framework for standardized, automated, and predictive radiotherapy in Finland

    by Jan Seppälä, Maria Tengström, Henri Korkalainen, Tuomas Virén, Juuso Tj Honkanen, Juha Nikkinen, Kaisa Lehtiö, Annika Ålgars, Jani Keyriläinen, Sami Suilamo, et al.

    Background and purpose: National radiotherapy (RT) data infrastructures are emerging to support treatment quality assurance and outcome research. However, prospective nationwide integration of full DICOM-RT data with toxicity and patient-reported outcome measures (PROMs) remains uncommon. KaSARi (Key advances in Standardizing, Automating, and predicting Risks in RT) was established to create national research infrastructure for prospective RT data collection and integration in Finland.Materials and methods: KaSARi is a prospective, consent-based multicenter cohort infrastructure involving Finnish university and central hospitals. All adult patients receiving RT are eligible. DICOM-RT datasets from treatment planning and oncology information systems are linked to clinician-graded toxicity, PROMs and relevant clinical variables. The infrastructure builds on previously validated multi-institutional DICOM workflows and follows FAIR (Findable, Accessible, Interoperable, Reusable) data principles. Predefined work packages support data harmonization, scalable data processing and enable downstream applications such as AI-based segmentation, dose prediction, and outcome modeling.Results: Patient recruitment started in May 2025. By February 2026, 221 patients had been enrolled at the coordinating center, initially including breast cancer patients and subsequently expanding to all RT indications. Recruitment at a second center began in November 2025, with 37 patients enrolled. The infrastructure is projected to include approximately 29,000 patients over 15 years. Integration of dosimetric, clinical and PROM data is progressing through a phased national implementation as additional centers join the network.Interpretation: KaSARi represents a national infrastructure prospectively integrating full DICOM-RT datasets with clinician-graded toxicity and PROMs across all RT indications within a unified research protocol. The study provides a foundation for nationwide outcome modeling, harmonization of RT practices and development and validation of data-driven AI-based methods.


  • Open Access

    Automated daily adaptation for breast cancer radiotherapy: benefits of deep learning-driven dose prediction workflow for CBCT-based adaptation

    by Nina Pesonen, Tuomas Virén, Mikko Mankinen, Akseli Leino, Janne Heikkilä, Hiran Cg Maladenige, Jan Seppälä, Henri Korkalainen

    Objective. Daily anatomical variations can jeopardize the quality of modern radiotherapy. Adaptive radiotherapy (ART) aims to address this by adjusting treatment plans according to the patient’s daily anatomy. However, as implementation of ART is currently time-consuming, laborious, and expensive, its routine clinical use remains limited. In this study, we investigate the feasibility of using deep learning (DL)-based dose prediction with automated treatment planning and daily cone-beam CTs (CBCTs) to evaluate the clinical benefits of daily replanning in breast cancer RT. Approach. Dataset of 28 breast cancer patients treated with Volumetric Modulated Arc Therapy and their daily CBCTs ( n = 15) were used to evaluate the dosimetric impact of daily anatomical variations. In-house developed DL-based dose prediction engine and automated planning solution were used to create dose accumulated and adaptive plans to determine the potential benefits of daily replanning. Main results. Adaptation improved the planning target volume (PTV) dose coverage in 25 of 28 patients and increased the fractions achieving the clinical goal (PTV V 38.05 Gy > 95%) from 35% to 93%. Moreover, ipsilateral lung V 1 6 Gy , heart V 4 Gy , and contralateral lung V 2 .5 Gy decreased by 0.4 pp, 0.6 pp, and 0.2 pp, respectively. Using an adaptation criterion based on DL prediction, approximately 25% of fractions were identified for replanning, resulting in a 1.2 pp increase in PTV V 3 8.05 Gy and reduced organs at risk (OAR) doses. Significance. This study demonstrated the feasibility of a CBCT-based automated workflow for breast cancer ART that combines DL-driven dose prediction with automated planning. The approach showed the benefit of replanning each fraction, improved PTV coverage, and offered potential OAR sparing, while also highlighting that daily adaptation may not benefit all patients on all fractions. With appropriately defined thresholds, the workflow can function as a decision-support tool, providing a practical framework to guide adaptive decision-making and optimize the use of clinical resources.


  • Open Access

    A comprehensive dose-volume histogram-based index for radiotherapy treatment plan quality evaluation: application to breast cancer radiotherapy

    by Hiran Cg Maladenige, Jan Seppälä, Tuomas Virén, Juuso Tj Honkanen, Janne Heikkilä, Akseli Leino, Henri Korkalainen

    Advances in radiotherapy have increased treatment plan complexity, making manual quality evaluation more subjective and variable. While deep learning approaches offer automation in planning, evaluation remains a manual bottleneck. Existing indices evaluate individual aspects in isolation and lack a unified perspective. To address this, we developed the advanced dose distribution index (ADDI), a comprehensive dose-volume histogram- and biological dose response-based metric for automation-friendly quality evaluation. ADDI summarizes treatment plan quality as a polar vector, ADDI( r , θ ), where the magnitude ( r ) reflects overall quality and the angle ( θ ) captures the balance between target and organ-at-risk dose. We validated ADDI across four breast cancer groups involving whole-breast irradiation with and without regional nodal irradiation for both lateralities, comprising 778 patients. ADDI demonstrated significant differences between groups, with left-sided and node-inclusive treatments having worse scores (i.e. higher dose exposure) than right-sided and node-exclusive cases. Additionally, ADDI responded to anatomical features such as planning target volume size, cardiothoracic ratio, central lung distance, and maximal heart distance, highlighting its sensitivity to patient-specific anatomical factors. These results indicate that ADDI effectively captures clinically meaningful variations in dose distribution and provides a unified, objective tool that is compatible with both automated and manual treatment planning workflows.

  • Open Access

    The benefit of DIBH on irradiated lung volume is overestimated in right-sided breast cancer VMAT with node involvement

    by Mikko Mankinen, Tuomas Virén, Jan Seppälä, Kristiina Vuolukka, Kimmo Myllyoja, Janne Heikkilä, Tuomas Koivumäki


  • 50 Hz magnetic field influences caspase-3 activity and cell cycle distribution in ionizing radiation exposed SH-SY5Y neuroblastoma cells

    by Valtteri Nieminen, Maria-Viola Martikainen, Saija Kalliomäki, Tuomas Virén, Jan Seppälä, Jukka Juutilainen, Jonne Naarala, Jukka Luukkonen

    PurposeEarlier evidence suggests that extremely low frequency magnetic fields (ELF MFs) can modify the effects of carcinogenic agents. However, the studies conducted so far with ionizing radiation as the co-exposure agent are sparse and have provided inconclusive results. We investigated whether 50 Hz MFs alone, or in combination with ionizing radiation alter cell biological variables relevant to cancer and the biological effects of ionizing radiation.Materials and methodsHuman SH-SY5Y neuroblastoma cells were sham exposed or exposed to 100 or 500 mu T MF for 24 h either before or after ionizing radiation exposure (0, 0.4 or 2 Gy). After the exposures, cells were assayed for viability, clonogenicity, reactive oxygen species, caspase-3 activity, and cell cycle distribution. Cell cycle distribution was assayed with propidium iodide staining followed by flow cytometry analysis and ROS levels were assayed together with cell viability by double staining with DeepRed and Sytox Blue followed by flow cytometry analysis.ResultsIncreased caspase-3 activity was observed in cells exposed to 500 mu T MF before or after ionizing radiation. Furthermore, exposure to the 500 mu T MF after the ionizing radiation decreased the percentage of cells in S-phase. No changes in the ROS levels, clonogenicity, or viability of the cells were observed in the MF exposed groups compared to the corresponding sham exposed groups, and no MF effects were observed in cells exposed at 100 mu T.ConclusionsOnly the 500 mu T magnetic flux density affected SH-SY5Y cells significantly. The effects were small but may nevertheless help to understand how MFs modify the effects of ionizing radiation. The increase in caspase-3 activity may not reflect effects on apoptosis, as no changes were observed in the subG1 phase of the cell cycle. In contrast to some earlier findings, 50 Hz MF exposure after ionizing radiation was not less effective than MF treatment given prior to ionizing radiation.

  • Open Access

    A Deep Learning-Based Automated CT Segmentation of Prostate Cancer Anatomy for Radiation Therapy Planning-A Retrospective Multicenter Study

    by Timo Kiljunen, Saad Akram, Jarkko Niemelä, Eliisa Löyttyniemi, Jan Seppälä, Janne Heikkilä, Kristiina Vuolukka, Okko-Sakari Kääriäinen, Vesa-Pekka Heikkilä, Kaisa Lehtiö, et al.

    A commercial deep learning (DL)-based automated segmentation tool (AST) for computed tomography (CT) is evaluated for accuracy and efficiency gain within prostate cancer patients. Thirty patients from six clinics were reviewed with manual- (MC), automated- (AC) and automated and edited (AEC) contouring methods. In the AEC group, created contours (prostate, seminal vesicles, bladder, rectum, femoral heads and penile bulb) were edited, whereas the MC group included empty datasets for MC. In one clinic, lymph node CTV delineations were evaluated for interobserver variability. Compared to MC, the mean time saved using the AST was 12 min for the whole data set (46%) and 12 min for the lymph node CTV (60%), respectively. The delineation consistency between MC and AEC groups according to the Dice similarity coefficient (DSC) improved from 0.78 to 0.94 for the whole data set and from 0.76 to 0.91 for the lymph nodes. The mean DSCs between MC and AC for all six clinics were 0.82 for prostate, 0.72 for seminal vesicles, 0.93 for bladder, 0.84 for rectum, 0.69 for femoral heads and 0.51 for penile bulb. This study proves that using a general DL-based AST for CT images saves time and improves consistency.


  • Breast deformation during the course of radiotherapy: The need for an additional outer margin

    by J. Seppälä, K. Vuolukka, T. Virén, J. Heikkilä, J.T.J. Honkanen, A. Pandey, A. Al-Gburi, M. Shah, S. Sefa, T. Koivumäki


  • Geometrical uncertainty of heart position in deep-inspiration breath-hold radiotherapy of left-sided breast cancer patients

    by Tuomas Koivumäki, Juuso Tujunen, Tuomas Virén, Janne Heikkilä, Jan Seppälä

    Background: This study aimed to determine the geometrical uncertainty of the position of the heart with daily cone beam computed tomography (CBCT) during deep-inspiration breath-hold (DIBH) treatment of the left breast. Material and methods: A visually guided optical respiratory monitoring system was used in DIBH treatment of 15 breast cancer patients. Heart position was determined in relation to the planning target volume (PTV) in 225 fractions in which daily low-dose CBCT images were compared with planning CT images. In addition, the position of the left lung apex and diaphragm was measured to evaluate the success of the DIBH. Results: The median shift of the heart was 1 mm to the left, 1 mm superiorly and 0 mm in the anterior–posterior (AP) direction during the treatment course when compared to the PTV position in planning CT. Based on these movements, an AP margin of 4 mm, a lateral (LR) margin of 3 mm, and a superior–inferior (SI) margin of 5 mm should be added to the heart contour to ensure avoiding the heart when planning treatment. The distance between the left lung apex and diaphragm, applied as a surrogate for lung volume, was 2mm (median) smaller during the CBCT acquisitions than during the planning CT acquisition. The correlation coefficient between the surrogate of lung volume and the distance between the heart and PTV was r = .46 in the AP, r = .72 in the LR and r = .79 in the SI directions. Conclusion: Residual variation was observed in the position of the heart in comparison to PTV, even with a visually guided DIBH technique. These geometrical uncertainties should be taken into account when planning radiotherapy treatment. The success of DIBH may make a major contribution to the variation of the heart position during treatment.