Tobias Hölscher
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Tobias Hölscher has a total of 120 co-authors. Here are the top co-authors by number of shared publications:
- Manfred P. Wirth
Universitätsklinkikum Carl Gustav Carus
6 shared publications - Fabian Lohaus
Department of Radiotherapy and Radiation Oncology, Faculty of Medicine and University Hospital Carl Gustav Carus TUD Dresden University of Technology Dresden Germany
4 shared publications - Jörg Kotzerke
Universitätsklinkikum Carl Gustav Carus
4 shared publications - Michael Baumann
European Academy of Cancer Sciences Stockholm Sweden
4 shared publications - Michael Froehner
Department of Urology, Zeisigwaldkliniken Bethanien Chemnitz, Zeisigwaldstrasse 101, D-09130 Chemnitz, Germany
4 shared publications - Steffen Löck
OncoRay - National Center for Radiation Research in Oncology, Faculty of Medicine and University Hospital Carl Gustav Carus, TUD Dresden University of Technology, Helmholtz-Zentrum Dresden-Rossendorf, Dresden, Germany
3 shared publications
Latest Publications
- Open Access
- 18 August 2025
- Published by Wiley in BJUI Compass
- Vol. 6 (8)
A machine learning algorithm for automatic tumour board recommendations in prostate cancer patients
Background and objective: Multidisciplinary tumour boards (MTBs) play a critical role in prostate cancer management, but their time‐intensive nature limits accessibility. This study evaluates machine learning (ML) algorithms for automating MTB recommendations in prostate cancer patients, focusing on multi‐label classification for diagnostic and therapeutic decisions.Methods: A retrospective dataset of 1929 MTB recommendations from 2020 to 2024 was used for model development and validation at a single academic centre. Three ML algorithms—Decision Tree, Random Forest and K‐Nearest Neighbours (KNN)—were trained to predict recommendations for PSMA‐PET, conventional imaging, active surveillance and local therapy (radical prostatectomy or radiotherapy). Model performance was assessed using accuracy, precision, recall and F1‐score.Key findings and limitations: The Random Forest model achieved the highest overall accuracy (66.3%, 95% CI 61.7–71%) and showed stable performance across most outcome categories. Predictions for local therapy were highly accurate (F1‐score: 0.99), but model performance was lower for less frequent recommendations such as PSMA‐PET and active surveillance, reflecting class imbalance and recent guideline changes. Limitations include moderate overall accuracy, retrospective single‐centre design and the need for extensive manual data preprocessing. In addition, a high proportion of patients were eligible for multiple treatment options, which may limit the discriminatory value of certain outcomes.Conclusions and clinical implications: This study demonstrates the potential of ML to replicate MTB decision patterns in prostate cancer with reasonable accuracy. However, the current model requires further optimization before it can be considered for clinical application. It should be regarded as a proof‐of‐concept that highlights both the opportunities and the challenges of algorithm‐based decision support in oncology. Future work should focus on improving model performance through multi‐institutional data, prospective validation and continuous adaptation to evolving clinical guidelines.
- 18 August 2025
- Published by Wiley in BJUI Compass
- Vol. 6 (8)
Open AccessA machine learning algorithm for automatic tumour board recommendations in prostate cancer patients
Background and objective: Multidisciplinary tumour boards (MTBs) play a critical role in prostate cancer management, but their time‐intensive nature limits accessibility. This study evaluates machine learning (ML) algorithms for automating MTB recommendations in prostate cancer patients, focusing on multi‐label classification for diagnostic and therapeutic decisions.Methods: A retrospective dataset of 1929 MTB recommendations from 2020 to 2024 was used for model development and validation at a single academic centre. Three ML algorithms—Decision Tree, Random Forest and K‐Nearest Neighbours (KNN)—were trained to predict recommendations for PSMA‐PET, conventional imaging, active surveillance and local therapy (radical prostatectomy or radiotherapy). Model performance was assessed using accuracy, precision, recall and F1‐score.Key findings and limitations: The Random Forest model achieved the highest overall accuracy (66.3%, 95% CI 61.7–71%) and showed stable performance across most outcome categories. Predictions for local therapy were highly accurate (F1‐score: 0.99), but model performance was lower for less frequent recommendations such as PSMA‐PET and active surveillance, reflecting class imbalance and recent guideline changes. Limitations include moderate overall accuracy, retrospective single‐centre design and the need for extensive manual data preprocessing. In addition, a high proportion of patients were eligible for multiple treatment options, which may limit the discriminatory value of certain outcomes.Conclusions and clinical implications: This study demonstrates the potential of ML to replicate MTB decision patterns in prostate cancer with reasonable accuracy. However, the current model requires further optimization before it can be considered for clinical application. It should be regarded as a proof‐of‐concept that highlights both the opportunities and the challenges of algorithm‐based decision support in oncology. Future work should focus on improving model performance through multi‐institutional data, prospective validation and continuous adaptation to evolving clinical guidelines.
- Open Access
- 13 February 2025
- Published by Elsevier in Annals of Oncology
- Vol. 36 (5) , 572-582
Predicting dose response to prostate cancer radiotherapy: validation of a radiation signature in the randomized phase III NRG/RTOG 0126 and SAKK 09/10 trials
- 13 February 2025
- Published by Elsevier in Annals of Oncology
- Vol. 36 (5) , 572-582
Open AccessPredicting dose response to prostate cancer radiotherapy: validation of a radiation signature in the randomized phase III NRG/RTOG 0126 and SAKK 09/10 trials
- 19 July 2023
- Published by Springer Nature
- p. 1571-1592
Prostatakarzinom: kurative Therapie
- 19 July 2023
- Published by Springer Nature
- p. 1593-1616
Prostatakarzinom: palliative Therapie
- 26 November 2022
- Published by Springer Nature
- p. 1-24
Prostatakarzinom: palliative Therapie
- Open Access
Local Control after Locally Ablative, Image-Guided Radiotherapy of Oligometastases Identified by Gallium-68-PSMA-Positron Emission Tomography in Castration-Sensitive Prostate Cancer Patients (OLI-P)
Progression of prostate-specific antigen (PSA) values after curative treatment of prostate cancer patients is common. Prostate-specific membrane antigen (PSMA-) PET imaging can identify patients with metachronous oligometastatic disease even at low PSA levels. Metastases-directed local ablative radiotherapy (aRT) has been shown to be a safe treatment option. In this prospective clinical trial, we evaluated local control and the pattern of tumor progression. Between 2014 and 2018, 63 patients received aRT of 89 metastases (MET) (68 lymph node (LN-)MET and 21 bony (OSS-)MET) with one of two radiation treatment schedules: 50 Gy in 2 Gy fractions in 34 MET or 30 Gy in 10 Gy fractions in 55 MET. The mean gross tumor volume and planning target volume were 2.2 and 14.9 mL, respectively. The median follow-up time was 40.7 months. Local progression occurred in seven MET, resulting in a local control rate of 93.5% after three years. Neither treatment schedule, target volume, nor type of lesion was associated with local progression. Regional progression in the proximity to the LN-MET was observed in 19 of 47 patients with at least one LN-MET (actuarial 59.3% free of regional progression after 3 years). In 33 patients (52%), a distant progression was reported. The median time to first tumor-related clinical event was 16.6 months, and 22.2% of patients had no tumor-related clinical event after three years. A total of 14 patients (22%) had another aRT. In conclusion, local ablative radiotherapy in patients with PSMA-PET staged oligometastatic prostate cancer may achieve local control, but regional or distant progression is common. Further studies are warranted, e.g., to define the optimal target volume coverage in LN-MET and OSS-MET.
Open AccessLocal Control after Locally Ablative, Image-Guided Radiotherapy of Oligometastases Identified by Gallium-68-PSMA-Positron Emission Tomography in Castration-Sensitive Prostate Cancer Patients (OLI-P)
Progression of prostate-specific antigen (PSA) values after curative treatment of prostate cancer patients is common. Prostate-specific membrane antigen (PSMA-) PET imaging can identify patients with metachronous oligometastatic disease even at low PSA levels. Metastases-directed local ablative radiotherapy (aRT) has been shown to be a safe treatment option. In this prospective clinical trial, we evaluated local control and the pattern of tumor progression. Between 2014 and 2018, 63 patients received aRT of 89 metastases (MET) (68 lymph node (LN-)MET and 21 bony (OSS-)MET) with one of two radiation treatment schedules: 50 Gy in 2 Gy fractions in 34 MET or 30 Gy in 10 Gy fractions in 55 MET. The mean gross tumor volume and planning target volume were 2.2 and 14.9 mL, respectively. The median follow-up time was 40.7 months. Local progression occurred in seven MET, resulting in a local control rate of 93.5% after three years. Neither treatment schedule, target volume, nor type of lesion was associated with local progression. Regional progression in the proximity to the LN-MET was observed in 19 of 47 patients with at least one LN-MET (actuarial 59.3% free of regional progression after 3 years). In 33 patients (52%), a distant progression was reported. The median time to first tumor-related clinical event was 16.6 months, and 22.2% of patients had no tumor-related clinical event after three years. A total of 14 patients (22%) had another aRT. In conclusion, local ablative radiotherapy in patients with PSMA-PET staged oligometastatic prostate cancer may achieve local control, but regional or distant progression is common. Further studies are warranted, e.g., to define the optimal target volume coverage in LN-MET and OSS-MET.
- 1 January 2022
- Published by Springer Nature
- p. 1-22
Prostatakarzinom: kurative Therapie
- Open Access
- 19 October 2021
- Published by Springer Nature in EJNMMI Research
- Vol. 11 (1) , 1-18
Intraindividual comparison of [68 Ga]-Ga-PSMA-11 and [18F]-F-PSMA-1007 in prostate cancer patients: a retrospective single-center analysis
Background: The analysis aimed to compare the radiotracers [68Ga]-Ga-PSMA-11 and [18F]-F-PSMA-1007 intraindividually in terms of malignant lesions, mi(molecular-imaging)TNM staging and presumable unspecific lesions retrospectively as used in routine clinical practice. Methods: A retrospective analysis of 46 prostate cancer patients (median age: 71 years) who underwent consecutive [68Ga]-Ga-PSMA-11- and [18F]-F-PSMA-1007-PET/CT or PET/MRI within a mean of 12 ± 8.0 days was performed. MiTNM staging was performed in both studies by two nuclear medicine physicians who were blinded to the results of the other tracer. After intradisciplinary and interdisciplinary consensus with two radiologists was reached, differences in both malignant and presumable nonspecific tracer accumulation were analyzed. Results: Differences in terms of miTNM stages in both studies occurred in nine of the 46 patients (19.6%). The miT stages differed in five patients (10.9%), the miN stages differed in three patients (6.5%), and different miM stages occurred only in one patient who was upstaged in [18F]-F-PSMA-1007 PET. Concordant miTNM stages were obtained in 37 patients (80.4%). There was no significant difference between [18F]-F-PSMA-1007 and [68Ga]-Ga-PSMA-11 in the SUVmax locally (31.5 vs. 32.7; p = 0.658), in lymph node metastases (28.9 vs. 24.9; p = 0.30) or in bone metastases (22.9 vs. 27.6; p = 0.286). In [18F]-F-PSMA-1007 PET, more patients featured presumable unspecific uptake in the lymph nodes (52.2% vs. 28.3%; p: < 0.001), bones (71.7% vs. 23.9%; p < 0.001) and ganglia (71.7% vs. 43.5%; p < 0.001). Probable unspecific, exclusively [18F]-F-PSMA-1007-positive lesions mainly occurred in the ribs (58.7%), axillary lymph nodes (39.1%) and cervical ganglia (28.3%). Conclusion: In terms of miTNM staging, both tracers appeared widely exchangeable, as no tracer relevantly outperformed the other. The differences between the two tracers were far more common in presumable unspecific lesions than in malignant spots. A routinely performed two-tracer study could not be shown to be superior. Since it seems at least challenging for most nuclear medicine departments to provide both [18F]-F-PSMA-1007 and [68Ga]-Ga-PSMA-11, it appears reasonable to choose the PSMA radiotracer depending on local availability with attention to the greater occurrence of nonspecific bone findings with [18F]-F-PSMA-1007.
- 19 October 2021
- Published by Springer Nature in EJNMMI Research
- Vol. 11 (1) , 1-18
Open AccessIntraindividual comparison of [68 Ga]-Ga-PSMA-11 and [18F]-F-PSMA-1007 in prostate cancer patients: a retrospective single-center analysis
Background: The analysis aimed to compare the radiotracers [68Ga]-Ga-PSMA-11 and [18F]-F-PSMA-1007 intraindividually in terms of malignant lesions, mi(molecular-imaging)TNM staging and presumable unspecific lesions retrospectively as used in routine clinical practice. Methods: A retrospective analysis of 46 prostate cancer patients (median age: 71 years) who underwent consecutive [68Ga]-Ga-PSMA-11- and [18F]-F-PSMA-1007-PET/CT or PET/MRI within a mean of 12 ± 8.0 days was performed. MiTNM staging was performed in both studies by two nuclear medicine physicians who were blinded to the results of the other tracer. After intradisciplinary and interdisciplinary consensus with two radiologists was reached, differences in both malignant and presumable nonspecific tracer accumulation were analyzed. Results: Differences in terms of miTNM stages in both studies occurred in nine of the 46 patients (19.6%). The miT stages differed in five patients (10.9%), the miN stages differed in three patients (6.5%), and different miM stages occurred only in one patient who was upstaged in [18F]-F-PSMA-1007 PET. Concordant miTNM stages were obtained in 37 patients (80.4%). There was no significant difference between [18F]-F-PSMA-1007 and [68Ga]-Ga-PSMA-11 in the SUVmax locally (31.5 vs. 32.7; p = 0.658), in lymph node metastases (28.9 vs. 24.9; p = 0.30) or in bone metastases (22.9 vs. 27.6; p = 0.286). In [18F]-F-PSMA-1007 PET, more patients featured presumable unspecific uptake in the lymph nodes (52.2% vs. 28.3%; p: < 0.001), bones (71.7% vs. 23.9%; p < 0.001) and ganglia (71.7% vs. 43.5%; p < 0.001). Probable unspecific, exclusively [18F]-F-PSMA-1007-positive lesions mainly occurred in the ribs (58.7%), axillary lymph nodes (39.1%) and cervical ganglia (28.3%). Conclusion: In terms of miTNM staging, both tracers appeared widely exchangeable, as no tracer relevantly outperformed the other. The differences between the two tracers were far more common in presumable unspecific lesions than in malignant spots. A routinely performed two-tracer study could not be shown to be superior. Since it seems at least challenging for most nuclear medicine departments to provide both [18F]-F-PSMA-1007 and [68Ga]-Ga-PSMA-11, it appears reasonable to choose the PSMA radiotracer depending on local availability with attention to the greater occurrence of nonspecific bone findings with [18F]-F-PSMA-1007.