J. Buijsen

‘Rapid Learning health care in oncology’ – An approach towards decision support systems enabling customised radiotherapy’ ☆ ☆☆

PURPOSEnAn overview of the Rapid Learning methodology, its results, and the potential impact on radiotherapy.nnnMATERIAL AND RESULTSnRapid Learning methodology is divided into four phases. In the data phase, diverse data are collected about past patients, treatments used, and outcomes. Innovative information technologies that support semantic interoperability enable distributed learning and data sharing without additional burden on health care professionals and without the need for data to leave the hospital. In the knowledge phase, prediction models are developed for new data and treatment outcomes by applying machine learning methods to data. In the application phase, this knowledge is applied in clinical practice via novel decision support systems or via extensions of existing models such as Tumour Control Probability models. In the evaluation phase, the predictability of treatment outcomes allows the new knowledge to be evaluated by comparing predicted and actual outcomes.nnnCONCLUSIONnPersonalised or tailored cancer therapy ensures not only that patients receive an optimal treatment, but also that the right resources are being used for the right patients. Rapid Learning approaches combined with evidence based medicine are expected to improve the predictability of outcome and radiotherapy is the ideal field to study the value of Rapid Learning. The next step will be to include patient preferences in the decision making.

Development and external validation of a predictive model for pathological complete response of rectal cancer patients including sequential PET-CT imaging

PURPOSEnTo develop and validate an accurate predictive model and a nomogram for pathologic complete response (pCR) after chemoradiotherapy (CRT) for rectal cancer based on clinical and sequential PET-CT data. Accurate prediction could enable more individualised surgical approaches, including less extensive resection or even a wait-and-see policy.nnnMETHODS AND MATERIALSnPopulation based databases from 953 patients were collected from four different institutes and divided into three groups: clinical factors (training: 677 patients, validation: 85 patients), pre-CRT PET-CT (training: 114 patients, validation: 37 patients) and post-CRT PET-CT (training: 107 patients, validation: 55 patients). A pCR was defined as ypT0N0 reported by pathology after surgery. The data were analysed using a linear multivariate classification model (support vector machine), and the models performance was evaluated using the area under the curve (AUC) of the receiver operating characteristic (ROC) curve.nnnRESULTSnThe occurrence rate of pCR in the datasets was between 15% and 31%. The model based on clinical variables (AUC(train)=0.61±0.03, AUC(validation)=0.69±0.08) resulted in the following predictors: cT- and cN-stage and tumour length. Addition of pre-CRT PET data did not result in a significantly higher performance (AUC(train)=0.68±0.08, AUC(validation)=0.68±0.10) and revealed maximal radioactive isotope uptake (SUV(max)) and tumour location as extra predictors. The best model achieved was based on the addition of post-CRT PET-data (AUC(train)=0.83±0.05, AUC(validation)=0.86±0.05) and included the following predictors: tumour length, post-CRT SUV(max) and relative change of SUV(max). This model performed significantly better than the clinical model (p(train)<0.001, p(validation)=0.056).nnnCONCLUSIONSnThe model and the nomogram developed based on clinical and sequential PET-CT data can accurately predict pCR, and can be used as a decision support tool for surgery after prospective validation.

FDG-PET provides the best correlation with the tumor specimen compared to MRI and CT in rectal cancer

PURPOSEnTo compare CT-, MR- and PET-CT based tumor length measurements in rectal cancer with pathology.nnnPATIENTS AND METHODSnTwenty-six rectal cancer patients underwent both MR and PET-CT imaging followed by short-course radiotherapy (RT 5×5 Gy) and surgery within 3 days after RT. Tumor length was measured manually and independently by 2 observers on CT, MR and PET. PET-based tumor length measurements were also generated automatically using the signal-to-background-ratio (SBR) method. All measurements were correlated with the tumor length on the pathological specimen.nnnRESULTSnCT-based measurements did not show a valuable correlation with pathology. MR-based measurements correlated only weakly, but still significantly (Pearson correlation=0.55 resp. 0.57; p<0.001). Manual PET measurements reached a good correlation with pathology, but less strong (Pearson correlation 0.72 and 0.76 for the two different observers) than automatic PET-CT based measurements, which provided the best correlation with pathology (Pearson correlation of 0.91 (p<0.001)). Bland-Altman analysis demonstrated in general an overestimation of the tumor diameter using manual measurements, while the agreement of automatic contours and pathology was within acceptable ranges. A direct comparison of the different modalities revealed a significant better precision for PET-based auto-contours as compared to all other measurements.nnnCONCLUSIONnAutomatically generated PET-CT based contours show the best correlation with the surgical specimen and thus provide a useful and powerful tool to accurately determine the largest tumor dimension in rectal cancer. This could be used as a quick and reliable tool for target delineation in radiotherapy. However, a 3D volume analysis is needed to confirm these results.

FDG–PET–CT reduces the interobserver variability in rectal tumor delineation

BACKGROUND AND PURPOSEnPreviously, we showed a good correlation between pathology and an automatically generated PET-contour in rectal cancer. This study analyzed the effect of the use of PET-CT scan on the interobserver variation in GTV definition in rectal cancer and the influence of PET-CT on treatment volumes.nnnMATERIALS AND METHODSnForty two patients diagnosed with rectal cancer underwent an FDG-PET-CT for radiotherapy planning. An automatic contour was created on PET-scan using the source-to-background ratio. The GTV was delineated by 5 observers in 3 rounds: using CT and MRI, using CT, MRI and PET and using CT, MRI and PET auto-contour. GTV volumes were compared and concordance indices (CI) were calculated. Since the GTV is only a small portion of the treatment volume in rectal cancer, a separate analysis was performed to evaluate the influence of PET on the definition of the CTV used in daily clinical practice and the caudal extension of the treatment volumes.nnnRESULTSnGTV volumes based on PET were significantly smaller. CIs increased significantly using PET and the best interobserver agreement was observed using PET auto-contours. Furthermore, we found that in up to 29% of patients the CTV based on PET extended outside the CTV used in clinical practice. The caudal border of the treatment volume can be tailored using PET-scan in low seated tumors. Influence of PET on the position of the caudal border was most pronounced in high seated tumors.nnnCONCLUSIONnPET-CT increases the interobserver agreement in the GTV definition in rectal cancer, helps to avoid geographical misses and allows tailoring the caudal border of the treatment volume.

Nomogram predicting response after chemoradiotherapy in rectal cancer using sequential PETCT imaging: a multicentric prospective study with external validation.

PURPOSEnTo develop and externally validate a predictive model for pathologic complete response (pCR) for locally advanced rectal cancer (LARC) based on clinical features and early sequential (18)F-FDG PETCT imaging.nnnMATERIALS AND METHODSnProspective data (i.a. THUNDER trial) were used to train (N=112, MAASTRO Clinic) and validate (N=78, Università Cattolica del S. Cuore) the model for pCR (ypT0N0). All patients received long-course chemoradiotherapy (CRT) and surgery. Clinical parameters were age, gender, clinical tumour (cT) stage and clinical nodal (cN) stage. PET parameters were SUVmax, SUVmean, metabolic tumour volume (MTV) and maximal tumour diameter, for which response indices between pre-treatment and intermediate scan were calculated. Using multivariate logistic regression, three probability groups for pCR were defined.nnnRESULTSnThe pCR rates were 21.4% (training) and 23.1% (validation). The selected predictive features for pCR were cT-stage, cN-stage, response index of SUVmean and maximal tumour diameter during treatment. The models performances (AUC) were 0.78 (training) and 0.70 (validation). The high probability group for pCR resulted in 100% correct predictions for training and 67% for validation. The model is available on the website www.predictcancer.org.nnnCONCLUSIONSnThe developed predictive model for pCR is accurate and externally validated. This model may assist in treatment decisions during CRT to select complete responders for a wait-and-see policy, good responders for extra RT boost and bad responders for additional chemotherapy.

Phase I trial of the combination of the Akt inhibitor nelfinavir and chemoradiation for locally advanced rectal cancer.

PURPOSEnTo investigate the toxicity of nelfinavir, administered during preoperative chemoradiotherapy (CRT) in patients with locally advanced rectal cancer.nnnMATERIAL AND METHODSnTwelve patients were treated with chemoradiotherapy to 50.4 Gy combined with capecitabine 825 mg/m(2) BID. Three dose levels (DL) of nelfinavir were tested: 750 mg BID (DL1), 1250 mg BID (DL2) and an intermediate level of 1000 mg BID (DL3). Surgery was performed between 8 and 10 weeks after completion of CRT. Primary endpoint was dose-limiting toxicity (DLT), defined as any grade 3 or higher non-hematological or grade 4 or higher hematological toxicity.nnnRESULTSnEleven patients could be analyzed: 5 were treated in DL1, 3 in DL2 and 3 in DL3. The first 3 patients in DL1 did not develop a DLT. In DL2 one patient developed gr 3 diarrhea, 1 patient had gr 3 transaminase elevation and 1 patient had a gr 3 cholangitis with unknown cause. An intermediate dose level was tested in DL3. In this group 2 patients developed gr 3 diarrhea and 1 patient gr 3 transaminase elevation and gr 4 post-operative wound complication. Three patients achieved a pathological complete response (pCR).nnnCONCLUSIONSnNelfinavir 750 mg BID was defined as the recommended phase II dose in combination with capecitabine and 50.4 Gy pre-operative radiotherapy in rectal cancer. First tumor response evaluations are promising, but a further phase II study is needed to get more information about efficacy of this treatment regimen.

Blood biomarkers are helpful in the prediction of response to chemoradiation in rectal cancer: A prospective, hypothesis driven study on patients with locally advanced rectal cancer

PURPOSE/OBJECTIVEnChemoradiation (CRT) has been shown to lead to downsizing of an important portion of rectal cancers. In order to tailor treatment at an earlier stage during treatment, predictive models are being developed. Adding blood biomarkers may be attractive for prediction, as they can be collected very easily and determined with excellent reproducibility in clinical practice. The hypothesis of this study was that blood biomarkers related to tumor load, hypoxia and inflammation can help to predict response to CRT in rectal cancer.nnnMATERIAL/METHODSn295 patients with locally advanced rectal cancer who were planned to undergo CRT were prospectively entered into a biobank protocol (NCT01067872). Blood samples were drawn before start of CRT. Nine biomarkers were selected, based on a previously defined hypothesis, and measured in a standardized way by a certified lab: CEA, CA19-9, LDH, CRP, IL-6, IL-8, CA IX, osteopontin and 25-OH-vitamin D. Outcome was analyzed in two ways: pCR vs. non-pCR and responders (defined as ypT0-2N0) vs. non-responders (all other ypTN stages).nnnRESULTSn276 patients could be analyzed. 20.7% developed a pCR and 47.1% were classified as responders. In univariate analysis CEA (p=0.001) and osteopontin (p=0.012) were significant predictors for pCR. Taking response as outcome CEA (p<0.001), IL-8 (p<0.001) and osteopontin (p=0.004) were significant predictors. In multivariate analysis CEA was the strongest predictor for pCR (OR 0.92, p=0.019) and CEA and IL-8 predicted for response (OR 0.97, p=0.029 and OR 0.94, p=0.036). The model based on biomarkers only had an AUC of 0.65 for pCR and 0.68 for response; the strongest model included clinical data, PET-data and biomarkers and had an AUC of 0.81 for pCR and 0.78 for response.nnnCONCLUSIONnCEA and IL-8 were identified as predictive biomarkers for tumor response and PCR after CRT in rectal cancer. Incorporation of these blood biomarkers leads to an additional accuracy of earlier developed prediction models using clinical variables and PET-information. The new model could help to an early adaptation of treatment in rectal cancer patients.

A systematic methodology review of phase I radiation dose escalation trials

BACKGROUND AND PURPOSEnThe purpose of this review is to evaluate the methodology used in published phase I radiotherapy (RT) dose escalation trials. A specific emphasis was placed on the frequency of reporting late complications as endpoint.nnnMATERIALS AND METHODSnWe performed a systematic literature review using a predefined search strategy to identify all phase I trials reporting on external radiotherapy dose escalation in cancer patients.nnnRESULTSnFifty-three trials (phase I: n = 36, phase I-II: n = 17) fulfilled the inclusion criteria. Of these, 20 used a modified Fibonacci design for the RT dose escalation, but 32 did not specify a design. Late toxicity was variously defined as > 3 months (n = 43) or > 6 months (n = 3) after RT, or not defined (n = 7). In only nine studies the maximum tolerated dose (MTD) was related to late toxicity, while only half the studies reported the minimum follow-up period for dose escalation (n = 26).nnnCONCLUSIONnIn phase I RT trials, late complications are often not taken into account and there is currently no consensus on the methodology used for radiation dose escalation studies. We therefore propose a decision-tree algorithm which depends on the endpoint selected and whether a validated early surrogate endpoint is available, in order to choose the most appropriate study design.

Image-guided stereotactic ablative radiotherapy for the liver: A safe and effective treatment

AIMSnStereotactic ablative body radiotherapy (SABR) is a non-invasive treatment option for inoperable patients or patients with irresectable liver tumors. Outcome and toxicity were evaluated retrospectively in this single-institution patient cohort.nnnPATIENTS AND METHODSnBetween 2010 and 2014, 39 lesions were irradiated in 33 consecutive patients (18 male, 15 female, median age of 68 years). All the lesions were liver metastases (n = 34) or primary hepatocellular carcinomas (n = 5). The patients had undergone four-dimensional respiration-correlated PET-CT for treatment simulation to capture tumor motion. We analyzed local control with a focus on CT-based response at three months, one year and two years after treatment, looking at overall survival and the progression pattern.nnnRESULTSnAll patients were treated with hypofractionated image-guided stereotactic radiotherapy. The equivalent dose in 2 Gy fractions varied from 62.5 Gy to 150 Gy, delivered in 3-10 fractions (median dose 93.8 Gy, alpha/beta = 10). The CT-based regression pattern three months after radiotherapy revealed partial regression in 72.7% of patients with a complete remission in 27.3% of the cases. The site of first progression was predominantly distant. One- and two-year overall survival rates were 85.4% and 68.8%, respectively. No toxicity of grade 2 or higher according to the NCI Common Terminology Criteria for Adverse Events v4.0 was observed.nnnCONCLUSIONnSABR is a safe and efficient treatment for selected inoperable patients or irresectable tumors of the liver. Future studies should combine SABR with systemic treatment acting in synergy with radiation, such as immunological interventions or hypoxic cell radiosensitizers to prevent distant relapse.

A phase I–II study on the combination of rapamycin and short course radiotherapy in rectal cancer

PURPOSEnThis phase I/II study sought to determine the safety and maximum tolerated dose (MTD) of the combination of rapamycin, an mTOR inhibitor, with short-course radiotherapy in rectal cancer patients. Antitumor activity, changes in metabolic activity and perfusion on imaging, and changes in phosphorylation status of the mTOR pathway were also assessed.nnnMATERIALS AND METHODSnPatients with primary resectable rectal cancer were treated with short-course hypofractionated radiotherapy (5×5 Gy) combined with oral rapamycin 1 week before and during radiotherapy, followed by surgical resection.nnnRESULTSnThirteen patients were entered in phase I. One patient developed a dose-limiting toxicity, consisting of a grade 4 leak and grade 4 bleeding. Because of an unexpected high rate of grade 3 postoperative toxicity, it was decided to treat patients with delayed surgery in phase II. Primary endpoint for phase II was tumor blood flow (K(trans)) assessed by perfusion CT. Thirty-one patients were treated with the MTD of 6 mg rapamycin daily. One patient (3%) developed a pathological complete response (pCR) and 3 patients (10%) had a ypT1N0 tumor at the time of resection. No change in tumor perfusion was observed on perfusion CT, but a significant decrease of metabolic activity was found on PET-scan.nnnCONCLUSIONSnThe combination of short-course radiotherapy and rapamycin turned out to be feasible, provided that the interval between neo-adjuvant treatment and surgical resection is at least 6 weeks. Although from this cohort no clear increase in pCR could be observed, a clear metabolic response after rapamycin run-in was observed, indicating a biological activity of this drug in rectal cancer.