It has become commonplace to use receiver operating curve (ROC) methodology to evaluate the incremental predictive accuracy of new markers in the presence of existing predictors. However, concerns have been raised about the validity of this practice. We have evaluated this issue in detail.
Simulations have been used that show clearly that use of risk predictors from nested models as data in subsequent tests comparing areas under the ROC curves of the models leads to grossly invalid inferences. Careful examination of the issue reveals two major problems: (1) the data elements are strongly correlated from case to case; and (2) the model that includes the additional marker has a tendency to interpret predictive contributions as positive information regardless of whether observed effect of the marker is negative or positive. Both of these phenomena lead to profound bias in the test.
We recommend strongly against the use of ROC methods derived from risk predictors from nested regression models to test the incremental information of a new marker.
18F-fluoromisonidazole PET, a noninvasive means of identifying hypoxia in tumors, has been widely applied but with mixed results, raising concerns about its accuracy. The objective of this study was to determine whether kinetic analysis of dynamic 18F-fluoromisonidazole data provides better discrimination of tumor hypoxia than methods based on a simple tissue-to-plasma ratio.
Eleven Dunning R3327-AT prostate tumor-bearing nude rats were immobilized in custom-fabricated whole-body molds, injected intravenously with 18F-fluoromisonidazole, and imaged dynamically for 105 min. They were then transferred to a robotic system for image-guided measurement of intratumoral partial pressure of oxygen (PO2). The dynamic 18F-fluoromisonidazole uptake data were fitted with 2 variants of a 2-compartment, 3-rate-constant model, one constrained to have K1 equal to k2 and the other unconstrained. Parametric images of the rate constants were generated. The PO2 measurements were compared with spatially registered maps of kinetic rate constants and tumor-to-plasma ratios.
The constrained pharmacokinetic model variant was shown to provide fits similar to that of the unconstrained model and did not introduce significant bias in the results. The trapping rate constant, k3, of the constrained model provided a better discrimination of low PO2 than the tissue-to-plasma ratio or the k3 of the unconstrained model.
The use of kinetic modeling on a voxelwise basis can identify tumor hypoxia with improved accuracy over simple tumor-to-plasma ratios. An effective means of controlling noise in the trapping rate constant, k3, without introducing significant bias, is to constrain K1 equal to k2 during the fitting process.
hypoxia; [18F]-FMISO; [18F]-fluoromisonidazole; kinetic modeling; parametric images; OxyLite probe
Nearly 25% of solid renal tumors are indolent cancer or benign and can be managed conservatively in selected patients. This prospective study was performed to determine whether preoperative IV microbubble contrast-enhanced ultrasound can be used to differentiate indolent and benign renal tumors from more aggressive clear cell carcinoma.
SUBJECTS AND METHODS
Thirty-four patients with renal tumors underwent preoperative gray-scale, color, power Doppler, and octafluoropropane microbubble IV contrast-enhanced ultrasound. Three blinded radiologists reading in consensus compared rate of contrast wash-in, grade and pattern of enhancement, and contrast washout compared with adjacent parenchyma. Contrast ultrasound findings were compared with surgical histopathologic findings for all patients.
The 34 patients had 23 clear cell carcinomas, three type 1 papillary carcinomas, one chromophobe carcinoma, one clear rare multilocular low-grade malignant tumor, two unclassified lesions, three oncocytomas, and one benign angiomyolipoma. The combination of heterogeneous lesion echotexture and delayed lesion washout had 85% positive predictive value, 43% negative predictive value, 48% sensitivity, and 82% specificity for predicting whether a lesion was conventional clear cell carcinoma or another tumor. Diminished lesion enhancement grade had 75% positive predictive value, 81% negative predictive value, 55% sensitivity, and 91% specificity for non–clear cell histologic features, either benign or low-grade malignant. Combining delayed washout with quantitative lesion peak intensity of at least 20% of kidney peak intensity had 91% positive predictive value, 40% negative predictive value, 63% sensitivity, and 80% specificity in the prediction of clear cell histologic features.
Ultrasound features of gray-scale heterogeneity, lesion washout, grade of contrast enhancement, and quantitative measure of peak intensity may be useful for differentiating clear cell carcinoma and non–clear cell renal tumors.
clear cell carcinoma; IV contrast medium; kidney; microbubbles; ultrasound
Historically, VHL was the only frequently mutated gene in clear cell renal cell carcinoma (ccRCC), with conflicting clinical relevance. Excitingly, recent sequencing efforts identified several novel, frequent mutations of histone modifying and chromatin remodeling genes in ccRCC, including PBRM1, SETD2, BAP1 and KDM5C. Intriguingly, PBRM1, SETD2 and BAP1 are located in close proximity to VHL within a commonly lost (~90%) 3p locus. To date the clinical and pathologic significance of mutations in these novel candidate tumor suppressors is unknown.
To determine the frequency of and render the first clinical and pathologic outcome associated with mutations of these novel candidate tumor suppressors in ccRCC.
Design, Setting, and Participants
Targeted sequencing was performed in 185 ccRCC and matched normal tissues from a single institute. Pathologic features, baseline patient characteristics and follow-up data were recorded.
The linkage between mutations and clinical and pathologic outcomes was interrogated with Fisher’s exact test (for stage and Fuhrman nuclear grade) and the permutation log-rank test (for cancer specific survival).
Results and Limitations
PBRM1, BAP1, SETD2 and KDM5C are mutated at 29%, 6%, 8% and 8%, respectively. Tumors with mutations in PBRM1 or any of BAP1, SETD2 or KDM5C (19%) are more likely to present with stage 3+ diseases, p=0.01 and p=0.001, respectively. Small tumors (<4cm) with PBRM1 mutations are more likely to exhibit stage 3 pathologic features (OR 6.4, p=0.001). BAP1 mutations tend to occur in Fuhrman Grade 3–4 tumors (p=0.052) and associate with worse cancer specific survival (p=0.01). Clinical outcome data is limited by the number of events.
Most mutations of chromatin modulators discovered in ccRCC are loss-of-function, which associate with advanced stage, grade, and possibly worsened cancer specific survival. Further studies validating the clinical impact of these novel mutations and future development of therapeutics remedying these tumor suppressors are warranted.
Chromatin; Histone; Mutation; Outcome; Renal Cell Carcinoma
Aging is characterized by clonal expansion of myeloid-biased hematopoietic stem cells and by an increased risk of myeloid malignancies. Exome sequencing of 3 elderly females with clonal hematopoiesis demonstrated by X-inactivation analysis identified somatic TET2 mutations. Recurrence testing found TET2 mutations in 10 out of 182 individuals with X-inactivation skewing. TET2 mutations were specific to individuals with clonal hematopoiesis without hematologic malignancies and were associated with alterations in DNA methylation.
TET2; X-inactivation; clonality; skewing; aging; clonal hematopoiesis; methylation; hydroxymethylation; HUMARA
The performance of prediction models can be assessed using a variety of different methods and metrics. Traditional measures for binary and survival outcomes include the Brier score to indicate overall model performance, the concordance (or c) statistic for discriminative ability (or area under the receiver operating characteristic (ROC) curve), and goodness-of-fit statistics for calibration.
Several new measures have recently been proposed that can be seen as refinements of discrimination measures, including variants of the c statistic for survival, reclassification tables, net reclassification improvement (NRI), and integrated discrimination improvement (IDI). Moreover, decision–analytic measures have been proposed, including decision curves to plot the net benefit achieved by making decisions based on model predictions.
We aimed to define the role of these relatively novel approaches in the evaluation of the performance of prediction models. For illustration we present a case study of predicting the presence of residual tumor versus benign tissue in patients with testicular cancer (n=544 for model development, n=273 for external validation).
We suggest that reporting discrimination and calibration will always be important for a prediction model. Decision-analytic measures should be reported if the predictive model is to be used for making clinical decisions. Other measures of performance may be warranted in specific applications, such as reclassification metrics to gain insight into the value of adding a novel predictor to an established model.
Pancreatic cancer is a leading cause of cancer-related death, largely due to metastatic dissemination. We investigated pancreatic cancer progression by utilizing a mathematical framework of metastasis formation together with comprehensive data of 228 patients, 101 of whom had autopsies. We found that pancreatic cancer growth is initially exponential. After estimating the rates of pancreatic cancer growth and dissemination, we determined that patients likely harbor metastases at diagnosis and predicted the number and size distribution of metastases as well as patient survival. These findings were validated in an independent database. Finally, we analyzed the effects of different treatment modalities, finding that therapies which efficiently reduce the growth rate of cells earlier in the course of treatment appear to be superior to upfront tumor resection. These predictions can be validated in the clinic. Our interdisciplinary approach provides insights into the dynamics of pancreatic cancer metastasis and identifies optimum therapeutic interventions.
Vorinostat (V) at levels >2.5 μM enhances chemotherapy in vitro. Yet the approved oral dose of 400 mg inconsistently achieves this level in patients. We developed an intermittent oral pulse-dose schedule of V to increase serum levels. We combined V with the cyclin dependent kinase inhibitor flavopiridol (F) which increases V-induced apoptosis.
One week before combination treatment, V alone was given daily for 3d (cycle −1). Then V was given on d1-3 and d8-10, and F on d2 and d9, every 21-d. Due to neutropenia, this was modified to V on d1-3 and d15–17, and F on d2 and d16, every 28-d. Bolus and split-dose F schedules were studied.
34 patients were treated. On the 21-d schedule, the maximum tolerated dose (MTD) was V 600 mg/d and F 60 mg/m2 bolus. On the 28-d schedule, the MTD was V 800 mg/d and F 30 mg/m2 over 30 min and 30 mg/m2 over 4 h. V Cmax at the 800 mg dose was 4.8 μM (± 2.8). V Cmax ≥2.5 μM was achieved in 86% of patients at the MTD. F increased the Cmax of V by 27% (95% CI 11%–43%). F Cmax of ≥2 μM was achieved in 90% of patients. 8 patients had stable disease for on average 5.5 m (range 1.6–13.2 m).
Intermittent high dose oral V in combination with F is feasible and achieves target serum levels >2.5 μM. V concentrations higher than previously reported with oral dosing were achieved.
CDKs and CDK inhibitors; Histone deacetylase inhibitors; Phase I trials; Combination chemotherapy; Pharmacokinetics
Flavopiridol, a Cdk inhibitor, potentiates irinotecan-induced apoptosis. In a phase I trial of sequential irinotecan and flavopiridol, 2 patients with advanced hepatocellular carcinoma (HCC) had stable disease (SD) for ≥14 months. We thus studied the sequential combination of irinotecan and flavopiridol in patients with HCC.
Patients with advanced HCC naïve to systemic therapy, Child-Pugh ≤B8, and Karnofsky performance score (KPS) ≥70% received 100 mg/m2 irinotecan followed 7 hours later by flavopiridol 60 mg/m2 weekly for 4 of 6 weeks. The primary end point was an improvement in progression-free survival at 4 months (PFS-4) from 33% to 54%, using a Simon's two-stage design. Tumors were stained for p53.
Only 16 patients in the first stage were enrolled: median age, 64 years; median KPS, 80%; Child-Pugh A, 87.5%; and stage III/IV, 25%/75%. The primary end point was not met; PFS-4 was 20%, leading to early termination of the study. Ten patients were evaluable for response: 1 had SD >1 year and 9 had disease progression. Grade 3 fatigue, dehydration, diarrhea, neutropenia with or without fever, lymphopenia, anemia, hyperbilirubinemia, and transaminitis occurred in ≥10% of the patients. Of the 9 patients who progressed, 5 had mutant p53 and 4 had wild-type p53. The patient with stable disease had wild-type p53.
Sequential irinotecan and flavopiridol are ineffective and poorly tolerated in patients with advanced HCC. Despite our limited assessments, it is possible that the presence of wild-type p53 is necessary but not sufficient to predict response in HCC.
Oncolytic viruses show promise for treating cancer. However, to assess therapy and potential toxicity, a noninvasive imaging modality is needed. This study aims to determine the in vivo biodistribution, and imaging and timing characteristics of a vaccinia virus, GLV-1h153, encoding the human sodium iodide symporter (hNIS.
GLV-1h153 was modified from GLV-1h68 to encode the hNIS gene. Timing of cellular uptake of radioiodide 131I in human pancreatic carcinoma cells PANC-1 was assessed using radiouptake assays. Viral biodistribution was determined in nude mice bearing PANC-1 xenografts, and infection in tumors confirmed histologically and optically via Green Fluorescent Protein (GFP) and bioluminescence. Timing characteristics of enhanced radiouptake in xenografts were assessed via 124I-positron emission tomography (PET). Detection of systemic administration of virus was investigated with both 124I-PET and 99m-technecium gamma-scintigraphy.
GLV-1h153 successfully facilitated time-dependent intracellular uptake of 131I in PANC-1 cells with a maximum uptake at 24 hours postinfection (P<0.05). In vivo, biodistribution profiles revealed persistence of virus in tumors 5 weeks postinjection at 109 plaque-forming unit (PFU)/gm tissue, with the virus mainly cleared from all other major organs. Tumor infection by GLV-1h153 was confirmed via optical imaging and histology. GLV-1h153 facilitated imaging virus replication in tumors via PET even at 8 hours post radiotracer injection, with a mean %ID/gm of 3.82±0.46 (P<0.05) 2 days after intratumoral administration of virus, confirmed via tissue radiouptake assays. One week post systemic administration, GLV-1h153-infected tumors were detected via 124I-PET and 99m-technecium-scintigraphy.
GLV-1h153 is a promising oncolytic agent against pancreatic cancer with a promising biosafety profile. GLV-1h153 facilitated time-dependent hNIS-specific radiouptake in pancreatic cancer cells, facilitating detection by PET with both intratumoral and systemic administration. Therefore, GLV-1h153 is a promising candidate for the noninvasive imaging of virotherapy and warrants further study into longterm monitoring of virotherapy and potential radiocombination therapies with this treatment and imaging modality.
Humanized A33 (huA33) is a promising monoclonal antibody that recognizes A33 antigen, which is present in more than 95% of colorectal cancers and in normal bowel. In this study, we took advantage of quantitative PET to evaluate 124I huA33 targeting, biodistribution, and safety in patients with colorectal cancer. We also determined the biodistribution of 124I-huA33 when a large dose of human intravenous IgG (IVIG) was administered to manipulate the Fc receptor or when 124I-huA33 was given via hepatic arterial infusion (HAI).
We studied 25 patients with primary or metastatic colorectal cancer; 19 patients had surgical exploration or resection. Patients received a median of 343 MBq (44.4–396 MBq) and 10 mg of 124I-huA33. Nineteen patients received the antibody intravenously and 6 patients via HAI, and 5 patients also received IVIG.
Ten of 12 primary tumors were visualized in 11 patients. The median concentration in primary colon tumors was 0.016% injected dose per gram, compared with 0.004% in normal colon. The PET-based median ratio of hepatic tumor uptake to normal-liver uptake was 3.9 (range, 1.8–22.2). Quantitation using PET, compared with well counting of serum and tissue, showed little difference. Prominent uptake in bowel hindered tumor identification in some patients. Pharmacokinetics showed that patients receiving IVIG had a significantly shorter serum half-time (41.6 ± 14.0 h) than those without (65.2 ± 9.8 h). There were no differences in clearance rates among the intravenous group, IVIG group, and HAI group, nor was there any difference in serum area under the curve, maximum serum concentration, or volume of distribution. Weak titers of human–anti-human antibodies were observed in 6 of 25 patients. No acute side effects or significant toxicities were associated with huA33.
Good localization of 124I-huA33 in colorectal cancer with no significant toxicity has been observed. PET-derived 124I concentrations agreed well with those obtained by well counting of surgically resected tissue and blood, confirming the quantitative accuracy of 124I-huA33 PET. The HAI route had no advantage over the intravenous route. No clinically significant changes in blood clearance were induced by IVIG.
A33; 124I; antibody; arterial; positron; colon
To evaluate the clinical outcomes of percutaneous radiofrequency (RF) ablation of colorectal cancer liver metastases (CLM) recurring after hepatectomy.
From December 2002 to December 2008 we ablated 71 CLM developing after hepatectomy in 56 patients. We reviewed medical records and imaging to determine: technique effectiveness/complete ablation (ablation defect covering the entire tumor on 4–6 week post-ablation CT), complications and local tumor progression at the site of ablation. Local tumor progression-free and overall survivals were calculated using Kaplan-Meier methodology. A modified clinical risk score (CRS) including nodal status of the primary, time interval from primary to liver metastases, number of tumors and size of the largest tumor was correlated to overall survival and local tumor progression.
Tumor size ranged between 0.5 and 5.7 cm. Complete ablation was documented in 67/71 (94%) CLM. Complications were: liver abscess (1) and pleural effusion (1). Median overall survival was 31 months. One-, 2- and 3-year overall survival rates were 91%, 66% and 41% respectively. CRS was an independent factor for overall survival (74% for CRS 0–2 vs. 42% for CRS 3–4 at 2 years p=0.03) and for local tumor progression-free survival (66% for CRS 0–2 vs 22% for CRS 3–4 at one year after a single ablation p<0.01).
CT-guided RF ablation can be used to treat recurrent CLM after hepatectomy. A low CRS is associated with better clinical outcomes.
Sphingosine 1-phosphate (S1P) is an important mediator of cancer cell growth and proliferation. Production of S1P is catalyzed by sphingosine kinase 1 (SphK). Safingol, (L-threo-dihydrosphingosine) is a putative inhibitor of SphK. We conducted a phase I trial of safingol (S) alone and in combination with cisplatin (C).
A 3+3 dose escalation was used. For safety, S was given alone 1 week before the combination. S + C were then administered every 3 weeks. S was given over 60–120 minutes (min), depending on dose. 60 min later, C was given over 60 min. The C dose of 75 mg/m2 was reduced in cohort 4 to 60 mg/m2 due to excessive fatigue.
43 patients were treated. 41 were evaluable for toxicity and 37 for response. The maximum tolerated dose (MTD) was S 840 mg/m2 over 120 min C 60 mg/m2, every 3 weeks. DLTs attributed to cisplatin included fatigue and hyponatremia. DLT from S was hepatic enzyme elevation. S pharmacokinetic parameters were linear throughout the dose range with no significant interaction with C. Patients treated at or near the MTD achieved S levels of > 20 µM and maintained levels ≥ 5 µM for 4 hours. The best response was stable disease in 6 patients for on average 3.3 months (range 1.8 – 7.2 m). One patient with adrenal cortical cancer had significant regression of liver and lung metastases and another had prolonged stable disease. S was associated with a dose-dependent reduction in S1P in plasma.
Safingol, the first putative SphK inhibitor to enter clinical trials, can be safely administered in combination with cisplatin. Reversible dose-dependent hepatic toxicity was seen, as expected from preclinical data. Target inhibition was achieved with downregulation of S1P. The recommended phase 2 dose is S 840 mg/m2 and C 60 mg/m2, every 3 weeks.
Drug-mediated stimulation of cell death pathways; Pharmacokinetics and pharmacodynamics; Kinase and phosphatase inhibitors; Novel antitumor agents; Sphingosine Kinase; Sphingolipids
Positive peritoneal fluid cytology predicts poor outcome in patients with resected pancreatic cancer. Reverse transcription–polymerase chain reaction (RT-PCR) has been proposed as a more sensitive means of detection of peritoneal micrometastases than conventional cytology. The clinical significance of RT-PCR positivity in the absence of other evidence of peritoneal disease is unknown. The purpose of the current study was to determine the outcome RT-PCR positive/cytology-negative patients who underwent potentially curative resection.
Peritoneal washings were collected prospectively from 115 patients with pancreatic cancer undergoing diagnostic laparoscopy at a single institution. Specimens were analyzed by a cytopathologist and by RT-PCR for carcinoembryonic antigen (CEA).
Of the 115 patients, 62 (54%) underwent R0 resection. Eleven of the 62 patients (18%) had peritoneal washings that were negative by conventional cytology but positive for CEA by RT-PCR. Those 11 patients experienced early peritoneal and overall disease recurrence versus those who were RT-PCR negative (P = 0.001, P = 0.003, respectively) independent of nodal status.
RT-PCR for CEA is a sensitive and specific method for the detection of clinically significant peritoneal micrometastases from pancreatic cancer and it might identify a subgroup of patients with otherwise negative findings at staging laparoscopy who might respond better to treatment other than primary surgical resection.
The liver is believed to promote tolerance, which may be beneficial due to its constant exposure to foreign Ags from the portal circulation. Although dendritic cells (DCs) are critical mediators of immune responses, little is known about human liver DCs. We compared freshly purified liver DCs from surgical specimens with autologous blood DCs. Liver and blood DCs were equally immature, but had distinct subset compositions. BDCA-1+ DCs represented the most prevalent liver DC subset, whereas the majority of peripheral blood DCs were CD16+. Upon TLR4 ligation, blood DCs secreted multiple proinflammatory cytokines, whereas liver DCs produced substantial amounts of IL-10. Liver DCs induced less proliferation of allogeneic T cells both in a primary MLR and after restimulation. Similarly, Ag-specific CD4+ T cells were less responsive to restimulation when initially stimulated by autologous liver DCs rather than blood DCs. In addition, liver DCs generated more suppressive CD4+ CD25+FoxP3+ T regulatory cells and IL-4-producing Th2 cells via an IL-10-dependent mechanism. Our findings are critical to understanding hepatic immunity and demonstrate that human liver DCs promote immunologic hyporesponsiveness that may contribute to hepatic tolerance.
The cyclin-dependent kinase inhibitor flavopiridol increases irinotecan- and fluorouracil-induced apoptosis. We conducted a phase I trial of FOLFIRI + flavopiridol in patients with advanced solid tumors.
FOLFIRI + flavopiridol were administered every 2 weeks. Based on sequence-dependent inhibition, flavopiridol was given 3 h after irinotecan but before 5-FU. Two maximum tolerated doses were determined, one with flavopiridol administered over 1 h, and one with flavopiridol split as a 30-min bolus followed by a 4-h infusion.
A total of 74 patients were enrolled and 63 were evaluable. The MTD with FOLFIRI was flavopiridol 80 mg/m2 over 1 h or 35 mg/m2 bolus + 35 mg/m2 over 4 h. Dose-limiting toxicities were diarrhea, fatigue, neutropenia, and neuropathy. Clinical activity included 2 partial responses in small bowel cancer and bladder cancer and 1 complete response in mucosal melanoma. Stable disease was seen in 22 patients. Pharmacokinetic studies showed increasing Cmax with increasing flavopiridol dose. Clinical benefit was correlated with the presence of wild-type p53. Of 25 patients with colorectal cancer, 11 had as best response SD for >3 m (median 6 m, range 4.2–15.4 m), despite failing ≥1 irinotecan-containing regimen.
Treatment with flavopiridol and FOLFIRI is a safe and effective regimen. Concentrations of flavopiridol that enhance the effects of FOLFIRI can be achieved. Clinical activity is encouraging and includes prolonged stable disease in patients with irinotecan-refractory colorectal cancer.
CDKs and CDK inhibitors; Phase I trials; Gastrointestinal cancers: colorectal; Combination chemotherapy; Pharmacokinetics and pharmacodynamics; Novel antitumor agents
Oncolytic viruses show promise for treating cancer. However, to assess therapeutic efficacy and potential toxicity, a noninvasive imaging modality is needed. This study aimed to determine if insertion of the human sodium iodide symporter (hNIS) cDNA as a marker for non-invasive imaging of virotherapy alters the replication and oncolytic capability of a novel vaccinia virus, GLV-1h153.
GLV-1h153 was modified from parental vaccinia virus GLV-1h68 to carry hNIS via homologous recombination. GLV-1h153 was tested against human pancreatic cancer cell line PANC-1 for replication via viral plaque assays and flow cytometry. Expression and transportation of hNIS in infected cells was evaluated using Westernblot and immunofluorescence. Intracellular uptake of radioiodide was assessed using radiouptake assays. Viral cytotoxicity and tumor regression of treated PANC-1tumor xenografts in nude mice was also determined. Finally, tumor radiouptake in xenografts was assessed via positron emission tomography (PET) utilizing carrier-free 124I radiotracer.
GLV-1h153 infected, replicated within, and killed PANC-1 cells as efficiently as GLV-1h68. GLV-1h153 provided dose-dependent levels of hNIS expression in infected cells. Immunofluorescence detected transport of the protein to the cell membrane prior to cell lysis, enhancing hNIS-specific radiouptake (P < 0.001). In vivo, GLV-1h153 was as safe and effective as GLV-1h68 in regressing pancreatic cancer xenografts (P < 0.001). Finally, intratumoral injection of GLV-1h153 facilitated imaging of virus replication in tumors via 124I-PET.
Insertion of the hNIS gene does not hinder replication or oncolytic capability of GLV-1h153, rendering this novel virus a promising new candidate for the noninvasive imaging and tracking of oncolytic viral therapy.
Multivariable prediction models have been shown to predict cancer outcomes more accurately than cancer stage. The effects on clinical management are unclear. We aimed to determine whether a published multivariable prediction model for bladder cancer (“bladder nomogram”) improves medical decision making, using referral for adjuvant chemotherapy as a model.
We analyzed data from an international cohort study of 4462 patients undergoing cystectomy without chemotherapy 1969 – 2004. The number of patients eligible for chemotherapy was determined using pathologic stage criteria (lymph node positive or stage pT3 or pT4), and for three cut-offs on the bladder nomogram (10%, 25% and 70% risk of recurrence with surgery alone). The number of recurrences was calculated by applying a relative risk reduction to eligible patients' baseline risk. Clinical net benefit was then calculated by combining recurrences and treatments, weighting the latter by a factor related to drug tolerability.
A nomogram cut-off outperformed pathologic stage for chemotherapy for every scenario of drug effectiveness and tolerability. For a drug with a relative risk of 0.80, where clinicians would treat no more than 20 patients to prevent one recurrence, use of the nomogram was equivalent to a strategy that resulted in 60 fewer chemotherapy treatments per 1000 patients without any increase in recurrence rates.
Referring cystectomy patients to adjuvant chemotherapy on the basis of a multivariable model is likely to lead to better patient outcomes than the use of pathological stage. Further research is warranted to evaluate the clinical effects of multivariable prediction models.
ladder cancer; adjuvant chemotherapy; prognosis; decision support; outcomes
Flavopiridol, a cyclin-dependent kinase inhibitor, has promising clinical activity when combined with chemotherapy. Preclinical data indicate that flavopiridol enhances oxaliplatin (OX)- and fluorouracil (5FU)-induced apoptosis in a sequence-dependent manner.
We conducted a phase I trial of flavopiridol + FOLFOX (folinic acid, 5FU, and OX) for advanced solid tumors. Flavopiridol was administered every two weeks with OX before 5FU, based on sequence-dependent growth inhibition. Flavopiridol pharmacokinetics and p53 status were evaluated.
Forty-eight patients were treated on study. With dose escalation of OX (85 mg/m2) and 5FU (2400 mg/m2), dose-limiting toxicities (DLT) included hyponatremia, thrombocytopenia, and neutropenia. 5FU was subsequently reduced to allow for dose escalation of flavopiridol. DLTs with escalation of flavopiridol were nausea, vomiting, and neutropenia. The maximum tolerated dose (MTD) was flavopiridol 70 mg/m2, oxaliplatin 85 mg/m2, and 5FU 1800 mg/m2 continuous infusion over 48 hours. Clinical activity was noted in platinum-refractory germ cell tumors (GCTs): 3 out of 9 (33%) evaluable patients demonstrated a partial response on imaging, and 7 out of 10 (70%) had a decline in serum tumor markers. Responses were also observed in pancreatic, gastric, and sweat gland tumors. Flavopiridol pharmacokinetics had significant interpatient variability. At the MTD, tumor samples were p53 mutant (>30% positive cells) for responders and p53 wild-type for non-responders.
Flavopiridol with FOLFOX is a safe and tolerable regimen. Promising clinical activity was seen across tumor types. Encouraging results in the platinum-refractory GCT population has prompted a phase II trial which is currently open for accrual.
flavopiridol; FOLFOX; germ cell tumor; solid tumor; refractory
J591, a monoclonal antibody that targets the external domain of the prostate specific membrane antigen (PSMA), has potential as an agent for radioimmunotherapy. A pilot trial was carried out in patients with prostate cancer using repetitive administrations of escalating masses of J591. An analysis was carried out to assess (1) lesion detectability by 111InJ591 gamma camera imaging compared to standard imaging methods and (2) the effect of increasing antibody mass on lesion detectability, biodistribution and dosimetry.
Fourteen patients with metastatic prostate cancer received escalating amounts (10, 25, 50 and 100 mg) of J591 in a series of administrations each separated by 3 weeks. All antibody administrations included a fixed amount of radiolabeled antibody 111In-DOTA-J591 (2mg of J591 labeled with 185MBq (5 mCi) of 111In via the chelating agent DOTA). Three whole body gamma camera scans with at least one SPECT scan together with multiple whole body count-rate measurements and serum activity concentration measurements were obtained in all patients. Images were analyzed for distribution and lesion targeting. Estimates of clearance rates and liver and lesion uptake were made for each treatment cycle. These estimates were used to generate dosimetric projections for radioimmunotherapy with 90Y-labeled J591.
A total of 80 lesions in 14 patients were detected. Both skeletal and soft tissue disease was targeted by the antibody as seen on 111In-J591 scans. Antibody localized to 93.7% of skeletal lesions detected by conventional imaging. Clearance of radioactivity from whole body, serum and liver was dependent on antibody mass. Normalized average values of the ratio of residence times between lesion and liver for 10, 25, 50 and 100mg of antibody were 1.0, 1.9, 3.2 and 4.0 respectively.
Dosimetric projections for radioimmunotherapy with 90Y-labeled J591 suggested similar absorbed doses to lesions, for treatment at the maximally tolerated activity (MTA), irrespective of antibody mass. However absorbed doses to liver at MTA would be antibody mass-dependent with estimates of 20, 10, 7 and 5 Gy for 10, 25, 50 and 100mg of J591respectively.
The proportion of the amount of antibody increased in lesions and decreased in the liver with increasing mass of administered antibody up to a dose of 50 mg. Proportional hepatic uptake continued to decrease with increasing antibody mass up to 100 mg. The optimal antibody mass for radioimmunotherapy would therefore appear to be greater than or equal to 50mg.
J591 antibody; prostate cancer; radioimmunotherapy; dosimetry
Medicare managed care may offer enrollees lower out-of-pocket costs and provide benefits that are not available in the traditional fee-for-service Medicare program. However, managed care plans may also restrict provider choice in an effort to control costs. We compared rates of voluntary disenrollment from Medicare managed care to traditional fee-for-service Medicare among Medicare managed care enrollees with and without a cancer diagnosis.
We identified Medicare managed care enrollees aged 65 years or older who were diagnosed with a first primary breast (n = 28 331), colorectal (n = 26 494), prostate (n = 29 046), or lung (n = 31 243) cancer from January 1, 1995, through December 31, 2002, in Surveillance, Epidemiology, and End Results (SEER) cancer registry records linked with Medicare enrollment files. Cancer patients were pair-matched to cancer-free enrollees by age, sex, race, and geographic location. We estimated rates of voluntary disenrollment to fee-for-service Medicare in the 2 years after each cancer patient’s diagnosis, adjusted for plan characteristics and Medicare managed care penetration, by use of Cox proportional hazards regression.
In the 2 years after diagnosis, cancer patients were less likely to disenroll from Medicare managed care than their matched cancer-free peers (for breast cancer, adjusted hazard ratio [HR] for disenrollment = 0.78, 95% confidence interval [CI] = 0.74 to 0.82; for colorectal cancer, HR = 0.84, 95% CI = 0.80 to 0.88; for prostate cancer, HR = 0.86, 95% CI = 0.82 to 0.90; and for lung cancer, HR = 0.81, 95% CI = 0.76 to 0.86). Results were consistent across strata of age, sex, race, SEER registry, and cancer stage.
A new cancer diagnosis between 1995 and 2002 did not precipitate voluntary disenrollment from Medicare managed care to traditional fee-for-service Medicare.
We previously demonstrated that there is a learning curve for open radical prostatectomy. We sought to determine whether the effects of the learning curve are modified by pathologic stage.
The study included 7765 eligible prostate cancer patients treated with open radical prostatectomy by one of 72 surgeons. Surgeon experience was coded as the total number of radical prostatectomies conducted by the surgeon prior to a patient’s surgery. Multivariable regression models of survival time were used to evaluate the association between surgeon experience and biochemical recurrence, with adjustment for PSA, stage, and grade. Analyses were conducted separately for patients with organ-confined and locally advanced disease.
Five-year recurrence-free probability for patients with organ-confined disease approached 100% for the most experienced surgeons. Conversely, the learning curve for patients with locally advanced disease reached a plateau at approximately 70%, suggesting that about a third of these patients cannot be cured by surgery alone.
Excellent rates of cancer control for patients with organ-confined disease treated by the most experienced surgeons suggest that the primary reason such patients recur is inadequate surgical technique.
Prostate cancer; Surgical learning curve; Decision analysis
Decision curve analysis is a novel method for evaluating diagnostic tests, prediction models and molecular markers. It combines the mathematical simplicity of accuracy measures, such as sensitivity and specificity, with the clinical applicability of decision analytic approaches. Most critically, decision curve analysis can be applied directly to a data set, and does not require the sort of external data on costs, benefits and preferences typically required by traditional decision analytic techniques.
In this paper we present several extensions to decision curve analysis including correction for overfit, confidence intervals, application to censored data (including competing risk) and calculation of decision curves directly from predicted probabilities. All of these extensions are based on straightforward methods that have previously been described in the literature for application to analogous statistical techniques.
Simulation studies showed that repeated 10-fold crossvalidation provided the best method for correcting a decision curve for overfit. The method for applying decision curves to censored data had little bias and coverage was excellent; for competing risk, decision curves were appropriately affected by the incidence of the competing risk and the association between the competing risk and the predictor of interest. Calculation of decision curves directly from predicted probabilities led to a smoothing of the decision curve.
Decision curve analysis can be easily extended to many of the applications common to performance measures for prediction models. Software to implement decision curve analysis is provided.
Oncolytic herpes simplex virus-1 (HSV-1) mutants selectively replicate in and lyse tumor cells. Viral replication is dependent on the cellular proliferative mechanism. Estrogen increases cellular proliferation and decreases apoptosis in estrogen receptor positive (ER+) human breast cancer cells. We hypothesize that the cellular changes produced by estrogen may enhance oncolytic viral replication and improve treatment of ER+ breast cancer cells.
Estrogen increased proliferation and replication of the HSV-1 mutant NV1066 in ER+ breast cancer cells. Additionally, cells grown with estrogen had lower rates of apoptosis and higher bcl-2 levels at baseline and after infection. Estrogen enhanced the oncolytic effect of NV1066, with cell kills of 95% and 97% at MOIs of 0.1 and 0.5, compared to 53% and 87% respectively without estrogen (p < 0.001).
Therapy of ER+ human breast cancer cells with a replication-competent HSV-1 mutant is improved in the presence of estrogen, in contrast to more standard therapies, such as chemotherapy and radiation, which demonstrate decreased efficacy in similar conditions. These data provide the mechanistic basis for the use of an oncolytic HSV-1 in patients with hormone receptor positive breast cancer, particularly if disease progresses with conventional therapies.
Apoptosis; Bcl-2; Herpes simplex; Hormonal therapy; Replication-competent viruses
Herpes simplex virus (HSV) oncolytic gene therapy is a promising treatment modality against cancer. We have demonstrated that androgen-induced cellular changes enhance oncolytic viral replication and improve efficacy in the treatment of androgen-dependent prostate cancer cell line. Imaging of changes in 18-F deoxyglucose (FDG) uptake by positron emission tomography (PET) is a sensitive method of detecting altered cellular metabolism involved in cancer therapy. We therefore hypothesized that FDG-PET can predict tumor response to oncolytic HSV therapy. In this study, androgen increased cell kill (74%) in vitro and enhanced viral yield (2.4-fold) in vivo following HSV therapy. This enhanced efficacy was predicted by high FDG accumulation in intact animals compared to low FDG uptake following orchiectomy (p=0.002). This proof-of-concept study provides the mechanistic basis for selecting patients for targeted oncolytic viral therapy by means of a noninvasive molecular imaging method in the treatment of prostate cancer.
Herpes virus; hormonal therapy; prediction; fluorodeoxyglucose