Abstract PARP inhibitors (PARPi) enhance radiation sensitivity in multiple cancer models, both in vitro and in vivo. Our observation that the radiosensitizing properties of PARPi are most pronounced in rapidly proliferating cells is reflected in early phase clinical trial data showing exacerbation of acute radiation toxicity in rapidly proliferating tissues such as oropharyngeal and esophageal mucosa. Lack of radiosensitization in late responding, slowly proliferating normal tissues indicates that PARPi may be more effectively combined with radiation therapy (RT) in patients with brain tumors. We are therefore evaluating the oral PARPi olaparib in combination with RT and/or temozolomide (TMZ) in the treatment of glioblastoma (GBM), the most prevalent and most aggressive primary brain tumor. Patients with GBM experience very poor outcomes in terms of median survival (c.1 year) and neurocognitive decline caused primarily by RT. Olaparib was initially evaluated in combination with daily low-dose TMZ in patients with recurrent GBM in the OPARATIC trial. Pharmacokinetic studies revealed that olaparib penetrates both core and margin regions of GBM, indicating that the BBB is significantly disrupted throughout these tumors. Olaparib could be safely combined with daily TMZ (75 mg/m2), but intermittent olaparib dosing (150 mg three days per week) was required to avoid dose-limiting hematological toxicity. Early phase testing of the olaparib-radiotherapy combination is now underway in three populations of patients with newly diagnosed GBM. Patients aged >65 with MGMT unmethylated GBM are being recruited to a randomized, placebo-controlled phase II study (PARADIGM) after a phase I dose escalation study showed that olaparib (200 mg twice daily) was extremely well tolerated when combined with brain irradiation (40 Gray in 15#). Good performance status patients aged <70 are being recruited to two parallel phase I dose escalation studies: patients with MGMT unmethylated tumors are receiving daily olaparib with RT (60 Gy in 30#) without TMZ, while patients with MGMT methylated tumors are receiving intermittent olaparib with standard chemoradiation (60 Gy). The impact of PARPi on RT induced neurotoxicity is being investigated in preclinical studies. In vitro data show that PARPi reduce proliferation of neural stem cells and protect them against RT induced apoptosis, while in vivo studies support the emerging concept that RT induced neuroinflammation is important in the pathogenesis of neurotoxicity. Importantly, preliminary PET and immunohistochemical studies have shown robust anti-neuroinflammatory effects of PARPi in this context. Ongoing experiments are defining the roles of microglia, astrocytes and neurogenesis in this phenomenon. These diverse data sets provide support for our hypothesis that combining PARPi with RT has potential to improve outcomes for GBM patients by enhancing tumor control while simultaneously suppressing neuroinflammation and alleviating RT related neurocognitive decline. Citation Format: Anthony J. Chalmers, Rodrigo Gutierrez-Quintana, David J. Walker, Karin Williams, Duncan Forster, Mark R. Jackson, Sarah Derby, Jon Stobo, Lorna Sweeting, Caroline Kelly, Stephen Durant, Kaye J. Williams. Enhancing the therapeutic ratio for glioblastoma by combining radiation therapy with PARP inhibitors [abstract]. In: Proceedings of the AACR Virtual Special Conference on Radiation Science and Medicine; 2021 Mar 2-3. Philadelphia (PA): AACR; Clin Cancer Res 2021;27(8_Suppl):Abstract nr IA-006.

Abstract Introduction: Pancreatic Cancer (PC) is the 3rd leading cause of cancer death. Precision-Panc UK Consortium was established to accelerate therapeutic development for PC by harmonizing the continuous learning between Discovery, Preclinical and Clinical Development, thereby overcoming the challenges of delivering precision medicine in PC. Methods: Central to the Clinical Development is a Master Protocol (MP) to screen, biopsy, and molecularly profile patients for subsequent enrollment into multiple downstream PRIMUS clinical trials on the portfolio. Novel patient and tissue pathways were developed and incorporated into routine clinical practice to fast-track research activities, to enable clinically meaningful turnaround time for molecular profile. All profiling is performed on bespoke Glasgow Precision Oncology Laboratory Clinical Cancer Genome assay using targeted capture next generation sequencing (NGS) technology. NGS is performed in batches for the first 150 patients on PRIMUS trial, then in real-time thereafter. Results: (Undated data to be presented if selected for presentation.) To date, 245 patients have entered the Precision-Panc study, of which 186 were registered on the MP, and 90 enrolled in PRIMUS trials. Patients not registered or subsequently not enrolled in PRIMUS trials were majority due to, alternative diagnosis (n = 16), declining performance status (n = 36), death (n = 21), or declined PRIMUS trials (n = 23). Of those declined, majority due to geography or wishing to start chemotherapy straight away. Two patients declined due to being germline BRCA carriers, and elected FOLFIRINOX. The median time between biopsy samples and DNA yield was 7 days (range 1 - 55). Of the first 100 donors confirmed to have PC on pathology QC, 21 underwent microdissection (cellularity < 15%), and 3 failed at DNA extraction QC (all had microdissection). Of the 97 donors with DNA (194 paired tumor/normal samples) submitted for sequencing (62 endoscopic ultrasound, 35 interventional radiology), 1/96 (1%) failed sequencing QC. KRAS mutation was not detected in 16 cases (17%), and out of these 8 did not have somatic variant and/or driver event identified indicating no tumor or low cellularity falling below the detection threshold of the sequencing and analysis pipeline. The other 8 cases with other driver events detected including actionable mutations such as NTRK1, and FGFR1 gain. In addition, 1 patient had TMB greater than 12 mut/MB. The NGS workflow is 10 working days turnaround. Each failed case was reviewed to identify causes and subsequent measures were made to prevent future occurrences. Conclusion: We present the initial Precision-Panc experience of delivering precision medicine for PC in a national health care system, highlighting challenges associated with attrition along patient and tissue pathways. Majority of samples, obtained by either EUS or IR biopsies, were successfully sequenced. Embedding research activities into routine clinical practice significantly reduced time from sample to DNA yield. Citation Format: David K. Chang, Susie Cooke, Stephan B. Dreyer, Jon Stobo, Fraser Duthie, Nigel Jamieson, Judith Dixon, Christine Wilshire, Nicola Williams, Colin J. McKay, Juan W. Valle, Andrew V. Biankin. Precision medicine for pancreatic cancer in national health care system: The initial Precision Panc experience [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 815.

e20054 Background: Lung cancer has poor outcomes, but those patients diagnosed with stage 1 NSCLC have 5-year survival of 60-70%. Stereotactic ablative body radiotherapy (SABR) is routinely considered for patients unfit for surgery. SABR is very well tolerated hence a treatment option for patients unfit for surgery, with > G2 lung toxicity of 1.8-9.1%, but propensity-matched analyses of stage 1 NSCLC treated with surgery or SABR have shown superiority of surgery in terms of overall survival, but not lung cancer-specific survival. Excess risk of CVD is a complication of breast cancer RT, but there is limited evidence in lung cancer. The aims of this study is to investigate the impact of radiation dose to the heart in lung SABR. Methods: In a single centre from November 2011-2012, patients who treated with lung SABR outwith the “no fly zone”, using ROSEL guidelines for OARS and with the lowest or highest heart maximum point dose were selected. 9 patients in each group were analysed for dose and clinical factors retrospectively from electronic patient notes. Results: The median (range) heart dose of the high dose (HD) group was 25.3Gy (17.6–34.1) and that of the low dose (LD) group was 0.88Gy (0.18–4.8). At the end of a median follow-up of 6.7 years, OS was poorer in the HD group at 11% than the LD group at 56% (p = 0.127). While 5-year cancer-specific mortality was similar in both groups at 22%, non-cancer-associated mortality was higher in HD (44% vs 11%, p = 0.127). In addition, HD were associated with a greater risk of CV hospital admission (56% vs 11%, p = 0.131) and new cardiac abnormalities (56% vs 0%, p = 0.029). Conclusions: In this group of medically unfit NSCLC patients treated with lung SABR, there were differences in outcome possibly associated with heart dose. The HD group showed higher mortality, specifically non-cancer-related mortality and cardiac events. Investigation of a larger cohort is warranted. [Table: see text]

Glioblastoma has a dismal prognosis and molecular targeted agents have failed to improve outcomes to date. PARADIGM-2 is a phase I dose escalation study evaluating olaparib plus radiotherapy ± temozolomide in newly diagnosed glioblastoma, using MGMT methylation status to stratify patients and inform treatment schedules.

Olaparib, a small molecule inhibitor of poly(ADP-ribose) polymerase (PARP), has radiosensitising properties in pre-clinical GBM models. Because radiopotentiation is observed only in proliferating cells, we hypothesised that olaparib would enhance tumour control without exacerbating normal brain toxicity in GBM patients receiving radiotherapy. Having shown that olaparib penetrates GBM at radiosensitising concentrations, we studied its safety and toxicity in combination with short-course radiotherapy in GBM patients ineligible for radical chemoradiation. Patients aged ≥70 (WHO PS 0–1) or <70 (PS 2) with histologically confirmed GBM received oral olaparib commencing three days before and continuing throughout radiotherapy (40 Gray in 15 fractions) and for four weeks afterwards. Olaparib dose was escalated in a 3 + 3 cohort design. 16 patients (9 male, 7 female) were treated within four olaparib dose cohorts. Median age was 72 (range 44–78); four patients had WHO PS 0, eight PS 1 and four PS 2. Cohort 3 was expanded to six evaluable patients because one patient experienced the only dose-limiting toxicity observed in the study (agitation grade 3, CTCAE v.4). Serious adverse events were experienced by eight patients, of which only one (the DLT) was a serious adverse reaction. The recommended dose of olaparib for phase II testing in combination with short-course radiotherapy was determined to be 200 mg twice daily. As predicted by pre-clinical data, olaparib is extremely well tolerated in combination with short-course radiotherapy in elderly and poorer PS patients with newly diagnosed GBM. The recommended phase II dose of 200 mg twice daily is significantly higher than has been deliverable to date in extracranial tumour sites in which acutely responding, rapidly proliferating normal tissues were within the irradiated volume. A randomised double-blind phase II study of radiotherapy plus olaparib versus radiotherapy plus placebo is underway in patients aged ≥65 with MGMT unmethylated GBM.

Lung cancer related mortality remains high after radiotherapy (RT) despite advances in treatment. The RTOG 0617 study (1) reported increased mortality within the higher radiation dose treatment arm, though toxicity was similar between the two study arms. One possible explanation was, increased radiation dose to the heart. Typically radiation induced heart disease (RIHD) is considered a late effect of RT in lymphoma and breast cancer. But RT dose prescribed in lung cancer is greater, and may result in acute effects in a population of patients with underlying cardio-pulmonary disease. Hence detailed dosimetric predictors are required for different cardiac morbidity endpoints. The CART study (2) was a prospective study investigating RIHD with serial cardiac MRI scans and the team involved developed a technique to analyse radiation dose of cardiac substructures without the MRI scan. CT planning scan was reformatted to develop a standardised method to outline in detail, cardiac substructures such as the left ventricular (LV) myocardium segments supplied by the Coronary Arteries (As). Other cardiac substructures such as cardiac chambers, conduction system and valves were contoured. This technique was applied to the planning scans of lung cancer patients and the dose to structures was calculated. Initially 5 patients who died within 3 months after radical RT were assessed. The RT treatment was 55Gy in 20 fractions over 4 weeks using 3D conformal RT technique. 3 patients had underlying cardiac or pulmonary comorbidities and pulmonary function as acceptable for all patients to proceed with radical RT. Dose to OARs were acceptable and all patients completed treatment. Radiation dose to the heart –mean heart dose range 158cGy-1910cGy, maximum heart dose range 1521-5669cGy and V20 dose range 0%-36% and V50 dose ranged from 0%-19%. Dose to left ventricular myocardium – LAD territory was maximum dose range (MaxDR) 71-2086 (cGy) and mean dose range (MeanDR) 112-802 (cGy); LCX territory was MaxDR 71-2549 (cGy) and MeanDR was 81-562 (cGy), and RCA territory was MaxDR 45-372 (cGy) and MeanDR was 44-178 (cGy) - demonstrated that anterior and lateral areas of LV myocardium received higher radiation dose. Dose to conduction system was high – SA node maximum dose ranged from 1140-5372 (cGy) and AV node maximum dose range was 103-1660 (cGy). It is feasible to use the CT planning scan to analyse retrospective patients for RT dose of coronary A myocardial territories, conduction system, and other substructures. An analysis of a larger sample of patients is planned.

Purpose Two previous single-arm trials have drawn conflicting conclusions regarding the activity of pazopanib in urothelial cancers after failure of platinum-based chemotherapy. Patients and Methods This randomized (1:1) open-label phase II trial compared the efficacy of pazopanib 800 mg orally with paclitaxel (80 mg/m2 days 1, 8, and 15 every 28 days) in the second-line setting. The primary end point was overall survival (OS). Results Between August 2012 and October 2014, 131 patients, out of 140 planned, were randomly assigned. The study was terminated early on the recommendation of the independent data monitoring committee because of futility. Final analysis after the preplanned number of deaths (n = 110) occurred after a median follow-up of 18 months. One hundred fifteen deaths had occurred at the final data extract presented here. Median OS was 8.0 months for paclitaxel (80% CI, 6.9 to 9.7 months) and 4.7 months for pazopanib (80% CI, 4.2 to 6.4 months). The hazard ratio (HR) adjusted for baseline stratification factors was 1.28 (80% CI, 0.99 to 1.67; one-sided P = .89). Median progression-free survival was 4.1 months for paclitaxel (80% CI, 3.0 to 5.6 months) and 3.1 months for pazopanib (80% CI, 2.7 to 4.6 months; HR, 1.09; 80% CI, 0.85 to 1.40; one-sided P = .67). Discontinuations for toxicity occurred in 7.8% and 23.1% for paclitaxel and pazopanib, respectively. Conclusion Pazopanib did not have greater efficacy than paclitaxel in the second-line treatment of urothelial cancers. There was a trend toward superior OS for paclitaxel.

Background There are limited data on the cardiac effects of radiotherapy and chemo-radiotherapy on the heart in patients with non-small cell lung cancer (NSCLC). CART is a pilot study designed to investigate change in myocardial function and tissue properties where Patients will undergo cardiac magnetic resonance (CMR) at baseline, during treatment, at 6 weeks and at 6 months after treatment completion. Here we report our preliminary findings related to temporal changes in myocardial perfusion index (MPI). Methods CMR was performed on a Siemens MAGNETOM Verio (Erlangen, Germany) 3.0 Tesla scanner. First-pass myocardial perfusion was assessed by saturation recovery prepared dynamic contrast enhanced sequence during administration of 0.1 mmol/kg of gadoterate meglumine (Dotarem). Semiquantitative analysis was performed on segmented basal and mid-LV short axis slices to derive normalised upslopes of myocardial signal intensity profiles (myocardial perfusion index, MPI). The change in MPI over time was analysed statistically using a linear mixed effects model. Results 13 patients currently have undergone CMR at baseline and during treatment (mean age 66 years, SD: 9; 71% male), 8 patients have undergone CMR at 6 weeks post treatment initiation (5 patients lost to follow-up/ died), and 6 patients have undergone CMR at 6 months from baseline (n = 2 lost to follow-up/ died). Perfusion index (Figure 1) varies significantly with time (p = 0.0015). After adjustment for multiple testing, the increase from baseline is statistically significant at 6 weeks (p = 0.014) and approaches significance at 6 months (p = 0.074) post treatment. Conclusion Change in microvascular perfusion is most pronounced 6 weeks following the completion of treatment. This finding will be used in the design of future clinical studies, where measurement of MPI will provide a robust comparison of different existing and emerging treatment protocols for NSCLC with regard to their effects on myocardial physiology. Funding Beatson Oncology Centre Fund Abstract 16 Figure 1 Boxplots of global myocardial perfusion index values by time point

INTRODUCTIONMalignant pleural mesothelioma (MPM) is associated with severe pain. The underlying neurobiology of this is complex. The primary aim of this study was to characterize pain in MPM.METHODSThis study was undertaken as part of a trial examining radiotherapy for the treatment of pain in MPM (ISRCTN 10644347). Patients had MPM with associated pain for which radiotherapy was planned and a worst pain score ≥ 4/10. The following assessments were undertaken: clinical neuropathic pain assessment, Brief Pain Inventory (BPI), Leeds Assessment of Neuropathic Symptoms and Signs (LANSS), Short form of the McGill Pain Questionnaire (SF-MPQ), and Quantitative Sensory Testing (QST). The relationship of these characteristics and response to radiotherapy was assessed. Unless stated, medians and interquartile range (IQR) are used.RESULTSThirty-seven patients were recruited. Average pain and worst pain was 4 (4-6) and 8 (6-8), respectively. Higher average pain and higher worst pain scores were associated with higher interference scores on the BPI, P < 0.001 and P < 0.0005. Twenty patients (54%) had a clinical diagnosis of neuropathic pain, and of these, only six patients (40%) screened positively for neuropathic pain using the LANSS. Patients with a high LANSS also had higher BPI and SF-MPQs. The presence of neuropathic pain (clinically or by LANSS) did not predict response to radiotherapy, P < 0.05. The SF-MPQ scores were higher in those with abnormal cool sensation on QST (P = 0.016).CONCLUSIONPain in mesothelioma varies among patients and may have neuropathic components. An adequate pain assessment is necessary to guide the clinician in the appropriate choice of analgesics.

Background:: Positron emission tomography computerized tomography (PET-CT) is useful in radiotherapy planning for lung cancer. However, its role in malignant pleural mesothelioma (MPM) is unknown. Objectives:: This exploratory study investigated the possible role for PET-CT in radiotherapy planning for MPM. Patients and Methods:: Patients receiving radiotherapy for the treatment of pain in MPM, had fluorodeoxyglucose (FDG) PET-CT scanning in addition to their standard CT scan. PET-CT images were then fused with CT planning images, termed Planning-PET-CT. Target volume delineation was undertaken first using CT and subsequently incorporated Planning-PET-CT. Planning treatment volume (PTV), conformity index (CI), mean distance to conformity (MDC), center of gravity distance (CGD) and standard uptake values (SUV) were examined. Results:: Sixteen patients were recruited into the study. PET-CT upstaged nine patients. No association between SUV max and either survival or pain response was seen. Volumes contoured using Planning-PET-CT differed markedly from those outlined using CT alone as shown by the following parameters: CI = 0.3 (0.24 - 0.38). MDC = 21.47 (16.73 - 33.70) and CGD = 16.40 (11.80 - 33.87). The median percentage of over contouring was 44.00% (34.33 - 72.50) with 46.67% under contoured (34.00 - 55.00). Conclusions:: PET-CT alters the position of the PTV in MPM and upstaged a proportion of patients. Further work to elucidate the role of Planning-PET-CT in target volume definition is justified.

Journal of Thoracic Oncology ® • Volume 10, Number 6, June 2015 Introduction: Radiotherapy is often used to treat pain in malignant pleural mesothelioma (MPM), although there is limited evidence to support this. The aim of this trial was to assess the role of radiotherapy for the treatment of pain in MPM. Methods: A multicentre, single arm phase II trial was conducted. Eligible patients fulfilled the following criteria: pathological or radiological diagnosis of MPM; pain secondary to MPM; radiotherapy indicated for pain control; and more than 18 years of age. Patients had assessments of pain and other symptoms at baseline and then received 20 Gy in five daily fractions. Key follow-up points were 5 and 12 weeks posttreatment. The primary end point measure was assessment of pain at the site of radiotherapy at 5 weeks. Secondary end points included effects on quality of life, breathlessness, fatigue, mood, toxicity, and the radiological response. Results: Forty patients were recruited from three UK oncology centers. Fourteen patients had a clinically meaningful improvement in their pain 5 weeks post radiotherapy (intention to treat), with five patients having a complete improvement. On the basis of a complete case analysis of the 30 patients assessable at week 5, 47% (confidence intervals, 28.3–65.7) of patients alive at week 5 had an improvement in their pain. There was no improvement in other key symptoms or quality of life. Conclusions: Radiotherapy for pain control in MPM is an effective treatment in a proportion of patients. Future studies examining differing radiotherapy regimens with a view to improving response rates are warranted.

Introduction: Radiotherapy is often used to treat pain in malignant pleural mesothelioma (MPM), although there is limited evidence to support this. The aim of this trial was to assess the role of radiotherapy for the treatment of pain in MPM.Methods: A multicentre, single arm phase II trial was conducted. Eligible patients fulfilled the following criteria: pathological or radiological diagnosis of MPM; pain secondary to MPM; radiotherapy indicated for pain control; and more than 18 years of age. Patients had assessments of pain and other symptoms at baseline and then received 20 Gy in five daily fractions. Key follow-up points were 5 and 12 weeks posttreatment. The primary end point measure was assessment of pain at the site of radiotherapy at 5 weeks. Secondary end points included effects on quality of life, breathlessness, fatigue, mood, toxicity, and the radiological response.Results: Forty patients were recruited from three UK oncology centers. Fourteen patients had a clinically meaningful improvement in their pain 5 weeks post radiotherapy (intention to treat), with five patients having a complete improvement. On the basis of a complete case analysis of the 30 patients assessable at week 5, 47% (confidence intervals, 28.3-65.7) of patients alive at week 5 had an improvement in their pain. There was no improvement in other key symptoms or quality of life.Conclusions: Radiotherapy for pain control in MPM is an effective treatment in a proportion of patients. Future studies examining differing radiotherapy regimens with a view to improving response rates are warranted.

Background Imatinib mesylate (IM) induces clinical remission of chronic myeloid leukemia (CML). The Abelson helper integration site 1 (AHI-1) oncoprotein interacts with BCR-ABL and Janus kinase 2 (JAK2) to mediate IM response of primitive CML cells, but the effect of the interaction complex on the response to ABL and JAK2 inhibitors is unknown. Methods The AHI-1–BCR-ABL–JAK2 interaction complex was analyzed by mutational analysis and coimmunoprecipitation. Roles of the complex in regulation of response or resistance to ABL and JAK2 inhibitors were investigated in BCR-ABL + cells and primary CML stem/progenitor cells and in immunodeficient NSG mice. All statistical tests were two-sided. Results The WD40-repeat domain of AHI-1 interacts with BCR-ABL, whereas the N-terminal region interacts with JAK2; loss of these interactions statistically significantly increased the IM sensitivity of CML cells. Disrupting this complex with a combination of IM and an orally bioavailable selective JAK2 inhibitor (TG101209 [TG]) statistically significantly induced death of AHI-1–overexpressing and IM-resistant cells in vitro and enhanced survival of leukemic mice, compared with single agents (combination vs TG alone: 63 vs 53 days, ratio = 0.84, 95% confidence interval [CI] = 0.6 to 1.1, P = .004; vs IM: 57 days, ratio = 0.9, 95% CI = 0.61 to 1.2, P = .003). Combination treatment also statistically significantly enhanced apoptosis of CD34+ leukemic stem/progenitor cells and eliminated their long-term leukemia-initiating activity in NSG mice. Importantly, this approach was effective against treatment-naive CML stem cells from patients who subsequently proved to be resistant to IM therapy. Conclusions Simultaneously targeting BCR-ABL and JAK2 activities in CML stem/progenitor cells may improve outcomes in patients destined to develop IM resistance.

Despite their efficacy in inducing deep and durable responses in chronic phase (CP) chronic myeloid leukaemia (CML) patients, BCR-ABL1 tyrosine kinase inhibitors (TKI) do not eradicate leukaemia stem cells (LSC), as proven by the persistence of BCR-ABL1+ CD34+, colony forming, long-term culture-initiating cells in the bone marrow of patients in sustained molecular response 4·5 (a 4·5-log reduction of BCR-ABL1 transcript levels, MR4·5) following TKI therapy (Chomel et al, 2011). CML LSC survival is independent of BCR-ABL1 kinase activity (Hamilton et al, 2012) and their quiescence is a putative TKI-resistance mechanism causing disease persistence. Moreover TKI exert anti-proliferative rather than pro-apoptotic effects against CML LSCs and might further contribute to disease persistence (Graham et al, 2002). Promoting LSC cell-cycle entry using granulocyte-colony stimulating factor (G-CSF) has been shown in vitro to restore their sensitivity to TKI and enhance their eradication (Jorgensen et al, 2006). Based on this evidence, we performed a randomized phase II study (GIMI, EudraCT 2004-000179-33), which compared the safety and efficacy of continuous imatinib (cIM) versus pulsed imatinib (pIM) alone or with G-CSF (pIM+G) therapy administered in 4-week cycles for 48 weeks (12 cycles in total) in CP CML patients with at least a complete cytogenetic response (CCyR) on IM (Drummond et al, 2009) (see reference for study design, primary and secondary endpoints, patient demographics and disease characteristics). The exp-erimental arms were expected to improve CML LSCs eradication by reducing TKI-induced quiescence (pIM) and/or by actively pushing CML LSCs into cell-cycle (pIM+G). At 2 years follow-up no statistically significant differences for the study endpoints were observed, possibly due to the limited numbers (15 patients per arm). However, 6/30 patients across the two experimental arms exhibited either loss of CCyR or major molecular response (MMR) as compared with only 1/15 in the cIM arm, with all but one patient in the experimental arms regaining MMR on restarting continuous IM or nilotinib therapy (Drummond et al, 2009). These results raised some concerns that the experimental schedules might have contributed to the loss of response. Subsequently, a mathematical model of the safety and efficacy of the IM and G-CSF combination suggested that this approach might be detrimental in the short- to medium-term for patients with persistent disease treated with IM, by increasing the LSCs burden through enhanced proliferation, thus in turn increasing the risk of acquiring a resistance mutation and of disease progression. However, in the long-term (>2500 d, i.e. 6·8 years, from start of treatment), such an approach was predicted to prove beneficial as it would deplete the CML LSCs by increasing their susceptibility to TKI (Foo et al, 2009). Here we report the 5-year follow-up data for the GIMI study. 41/45 patients were available for analysis; four patients had died (one only as a result of CML progression). The median follow-up was 5·67 years. Using an intention to treat analysis, both CCyR and MMR rates were similar among treatment arms with no differences in progression rates. 5/15 patients in the cIM arm compared to 3/15 patients in each experimental arm changed treatment to second generation TKI (Table 1). Of note, the reduction in BCR-ABL1 transcript levels between trial entry (baseline) and 5 years was highly statistically significant in both the pIM (1·10 median log reduction, P = 0·013) and the pIM+G arms (1·43 median log reduction, P = 0·002), while not significant in the cIM arm (1·06 median log reduction, P = 0·060) (Fig 1). However, a comparison of the change from baseline between treatment arms shows that, although this was lower in the cIM arm, it was not significantly different from the change in either the pIM (P = 0·678) or pIM+G arm (P = 0·528). Although the small number of patients demands caution in interpreting these data, these findings are reassuring regarding the safety of the experimental therapeutic approaches. Moreover, the deeper and significant reduction in BCR-ABL1 transcript levels in the experimental arms compared to the cIM arm at 5 years provocatively suggests that in the long-term, as suggested by the mathematical model, these treatment strategies might become beneficial to CP CML patients with persistent disease. However caution is required when making inferences based on the mathematical model of IM and G-CSF treatment, as it predicted the effects of an indefinite duration of this treatment strategy while in our study the combination treatment was only continued for 48 weeks. Nevertheless, considering these findings, we believe that a therapeutic strategy aiming to reverse LSC quiescence in CML combining IM and G-CSF could be safely pursued and only larger studies might be able to provide a definitive answer on its efficacy. This research was supported by the Leukaemia Research Fund CTAP (UK), Chugai Pharmaceutical Co. Ltd, Chugai Pharma, France and Novartis Pharmaceuticals (UK). We are grateful to Lindsay Mitchell and Karen Stewart for their help with recruitment and assessment of patients. We thank the ‘Délégation à la recherche clinique des Hospices Civils de Lyon’ for their support in setting up this trial in France, Elodie Gadolet and Madeleine Etienne, CRAs, for their help in collecting the data for the French patients. Drs Steve O'Brien, Heather Jorgensen and Alex McMahon assisted in reviewing the data. This study was supported by the Glasgow Experimental Cancer Medicine Centre, which is funded by Cancer Research UK and by the Chief Scientist's Office, Scotland. P.G. was funded by Medical Research Council UK clinical research training fellowship grant G1000288, N.H. was funded by the Leukaemia & Lymphoma Research UK clinical research training fellowship grant 05086 and T.L.H. was funded by Cancer Research UK Programme grant C11074/A11008. PG analysed data and wrote the manuscript, JS TH and JP analysed data and reviewed the manuscript, NH, FEN, RC, JT, LM, and MD looked after patients, collected data and reviewed the manuscript, GW collected data, FM reviewed the manuscript and TLH designed trial, looked after patients, collected/analysed data and reviewed manuscript. The authors declare no competing financial interests.