Michael Cerny Green

The role of anthracyclines in the treatment of gastric cancer

Prior to 1974 gastric cancer was considered refractory to chemotherapy. Single agent 5-Fluorouracil, mitomycin, and the nitrosoureas produced modest response rates with no augmentation of survival as compared to historical controls. Combinations of these agents produced slight increases in survival. The addition of doxorubicin, which had no major impact as a single agent, to 5-Fluorouracil and mitomycin C in 1974 produced impressive response rates over 40%, although these early optimistic results were tempered in randomized multi-institutional trials. Nevertheless, median overall survivals appear to have increased with this regimen over historical controls, and the substitution of cisplatin or methyl-CCNU for mitomycin C in Phase II trials has produced equivalent results. Comparison with historical controls is problematic as it is unknown what role more aggressive surgery and supportive measures may have played in increasing survival. Encouraging results with combination therapy in advanced gastric cancer have suggested opportunities to employ combination chemotherapy as adjuvant therapy or together with radiation for locally advanced gastric cancer. The role of less cardiotoxic derivatives of doxorubicin remains to be more fully explored.

Epirubicin has modest single-agent activity in head and neck cancer but limited activity in metastatic melanoma and colorectal cancer: Phase II studies by the Eastern Cooperative Oncology Group

Epirubicin (4-epidoxorubicin), a diastereoisomer of doxorubicin, has established activity in the treatment of many cancer types sensitive to doxorubicin. Its activity in other tumor targets such as melanoma, head and neck cancer, and recurrent colorectal cancer has been less well defined. Three concurrent phase II studies examined the efficacy and toxicity of epirubicin (90 mg/m2 given intravenously at 3-week intervals) in the treatment of 71 patients with the aforementioned cancers. Of 66 eligible patients who were assessable for response, one patient (with colorectal cancer) achieved a complete response and three patients (with head and neck cancer) achieved partial responses. The response rate in patients with head and neck cancer was 18% (95% confidence interval, 4-43%). Myelosuppression, alopecia, and nausea were the most frequent toxicities. Two patients died of neutropenic sepsis and grade IV leukopenia occurred in six patients (8%). Grade III toxicities were as follows: leukopenia (17%), anemia (10%), alopecia (8%), fever (1%), thrombocytopenia (1%). Grade I or II cardiac toxicity was noted in four patients at cumulative doses ranging between 375 mg/m2 to 1,283 mg/m2. Epirubicin is ineffective as a single agent at this dose and schedule in the treatment of patients with melanoma and colorectal cancer. In head and neck cancer, a modest response rate encourages further exploration of epirubicin and related anthracyclines in combination regimens.

Phase II study of carboplatin in recurrent ovarian cancer: severe hematologic toxicity in previously treated patients

SummaryCarboplatin (CBDCA) is a second-generation cisplatin analog that has shown activity in early clinical trials. Its spectrum of toxicity is quantitatively and qualitatively different from that of the parent compound. Between November 1984 and September 1986 we conducted a phase II trial of CBDCA in 46 women with epithelial ovarian cancer. All patients had undergone at least one prior chemotherapy regimen; 41 (89%) had previously received cisplatin (mean cumulative dose, 540 mg/m2). The CBDCA dose was based on renal function and was injected i. v. once every 4 weeks. Patients were stratified on the basis of baseline creatinine clearance: those with a baseline creatinine clearance of ≥60 ml/min received 400 mg/m2 CBDCA; those with a creatinine clearance between 30 and 60 ml/min received an initial dose calculated according to a previously published formula [2, 3] that corrected for renal insufficiency and projected nadir platelet counts of 75,000/mm3. Of 41 evaluable patients, 6 (15%) had an objective response [2 complete responses (CRs); 4 partial responses (PRs)]; 5 of the 6 responders had previously responded to cisplatin treatment. No responses were observed in 12 patients who had not responded to prior cisplatin therapy. Significant hematologic toxicity was seen. Of 18 patients with a creatinine clearance of ≥60 ml/min (dose, 400 mg/m2), 6 had nadir platelet counts of <25,000/mm3, 4 with symptomatic bleeding. Of the 21 evaluable patients for whom the dose-modification formula was applied, 10 had nadir platelet counts of <75,000/mm3; 5 had counts of <50,000/mm3. CBDCA has activity even in patients who have previously undergone extensive cisplatin therapy; however, its toxicity is variable and thrombocytopenia is dose-limiting. We did not confirm the ability of the above-mentioned formula to calculate the CBDCA dose and accurately predict the nadir platelet count for all patients. Other factors, such as prior radiotherapy, may also be important in the dosing of CBDCA in pretreated patients.

General video game rule generation

We introduce the General Video Game Rule Generation problem, and the eponymous software framework which will be used in a new track of the General Video Game AI (GVGAI) competition. The problem is, given a game level as input, to generate the rules of a game that fits that level. This can be seen as the inverse of the General Video Game Level Generation problem. Conceptualizing these two problems as separate helps breaking the very hard problem of generating complete games into smaller, more manageable subproblems. The proposed framework builds on the GVGAI software and thus asks the rule generator for rules defined in the Video Game Description Language. We describe the API, and three different rule generators: a random, a constructive and a search- based generator. Early results indicate that the constructive generator generates playable and somewhat interesting game rules but has a limited expressive range, whereas the search- based generator generates remarkably diverse rulesets, but with an uneven quality.

AtDELFI: automatically designing legible, full instructions for games

This paper introduces a fully automatic method for generating video game tutorials. The AtDELFI system (Automatically DEsigning Legible, Full Instructions for games) was created to investigate procedural generation of instructions that teach players how to play video games. We present a representation of game rules and mechanics using a graph system as well as a tutorial generation method that uses said graph representation. We demonstrate the concept by testing it on games within the General Video Game Artificial Intelligence (GVG-AI) framework; the paper discusses tutorials generated for eight different games. Our findings suggest that a graph representation scheme works well for simple arcade style games such as Space Invaders and Pacman, but it appears that tutorials for more complex games might require higher-level understanding of the game than just single mechanics.

DATA agent

This paper introduces DATA Agent, a system which creates murder mystery adventures from open data. In the game, the player takes on the role of a detective tasked with finding the culprit of a murder. All characters, places, and items in DATA Agent games are generated using open data as source content. The paper discusses the general game design and user interface of DATA Agent, and provides details on the generative algorithms which transform linked data into different game objects. Findings from a user study with 30 participants playing through two games of DATA Agent show that the game is easy and fun to play, and that the mysteries it generates are straightforward to solve.

Esorubicin (deoxydoxorubicin) has low grade activity in malignant melanoma

SummaryIn this phase II trial, twenty patients with advanced, measurable melanoma from ECOG institutions were treated with esorubicin 30 mg/m2 iv every three weeks. Doses were escalated or reduced based on nadir counts. The dose limiting toxicity was leukopenia with no significant thrombocytopenia or anemia. Other toxicities were mild. One patient had skin necrosis with extravasation. Two patients with soft tissue disease had partial remissions and were treated with 9 and 17 courses. One patient was stable for 8 courses. No cardiac toxicity was seen in three patients receiving more than 150 mg/m2. The response rate was 10% (90% CI = 2 to 30%). Low level activity was seen, but it is unlikely that this drug has sufficient activity to warrant further development in melanoma.

Generative design in minecraft (GDMC): settlement generation competition

This paper introduces the settlement generation competition for Minecraft, the first part of the Generative Design in Minecraft challenge. The settlement generation competition is about creating Artificial Intelligence (AI) agents that can produce functional, aesthetically appealing and believable settlements adapted to a given Minecraft map---ideally at a level that can compete with human created designs. The aim of the competition is to advance procedural content generation for games, especially in overcoming the challenges of adaptive and holistic PCG. The paper introduces the technical details of the challenge, but mostly focuses on what challenges this competition provides and why they are scientifically relevant.

Generating levels that teach mechanics

The automatic generation of game tutorials is a challenging AI problem. While it is possible to generate annotations and instructions that explain to the player how the game is played, this paper focuses on generating a gameplay experience that introduces the player to a game mechanic. It evolves small levels for the Mario AI Framework that can only be beaten by an agent that knows how to perform specific actions in the game. It uses variations of a perfect A* agent that are limited in various ways, such as not being able to jump high or see enemies, to test how failing to do certain actions can stop the player from beating the level.