Yapay Zekâ Sohbet Robotlarının Çocukluk Çağı Kanserlerinde Radyonüklid Tedavileriyle İlgili Yanıtları Ebeveynler İçin Güvenilir ve Okunabilir Kaynaklar mı?

Öz

AMAÇ: İki popüler yapay zekâ sohbet robotunun çocukluk çağı kanserleri için en yaygın radyonüklid tedavileri hakkında ebeveynlerin sorularına güvenilir ve okunabilir bilgi sağlayıp sağlamayacağını belirlenmesi amaçlandı. GEREÇ VE YÖNTEM: Tiroid kanseri için I-131 (RAI) ve nöroblastom için I-131 MIBG ile ilgili 35 ebeveyn sorusu, Kasım 2024'te Gemini Advanced ve ChatGPT-4'e ayrı ayrı soruldu. Yanıtlar, DISCERN ölçeği kullanılarak güvenilirlik açısından ve Gunning Fog Index (GFI), Flesch Reading Ease (FRE) ve Flesch-Kincaid Reading Grade Level (FKGL) kullanılarak okunabilirlik açısından değerlendirildi.Üç hakemden (bir nükleer tıp doktoru, ChatGPT-4 ve Gemini) elde edilen DISCERN puanlarının değerlendiriciler arası güvenilirliği analiz edildi. BULGULAR: Hekim tarafından RAI ve MIBG tedavileri hakkındaki yanıtlar için değerlendirilen DISCERN puanları (ortalama ± standart sapma) ChatGPT-4 için sırasıyla 54 ± 6.37 ve 60.8 ± 3.81 ve Gemini için 54.63 ± 6.01 ve 54.77 ± 5.12 idi. ChatGPT-4 tarafından RAI ve MIBG tedavileri hakkındaki cevaplar için elde edilen DISCERN skorları sırasıyla ChatGPT-4 için 50.26 ± 5.87 ve 58.09 ± 4.42 ve Gemini için 59.66 ± 6.91 ve 62.63 ± 4.88 idi. RAI ve MIBG ile ilgili yanıtlar için Gemini tarafından elde edilen DISCERN skorları ChatGPT-4 için 50.86 ± 6.33ve 54.09 ± 5.99; Gemini için 57.2 ± 6.97 ve 56.11 ± 7.29'dur. ChatGPT-4'ün RAI ve MIBG ile ilgili yanıtları için FKGL değerleri sırasıyla 13.34 ± 2.01 ve 12.8 ± 2.17 idi. Gemini'nin RAI ve MIBG ile ilgili yanıtları için FKGL değerleri sırasıyla 12.18 ± 1.47 ve 11.53 ± 1.74 idi. SONUÇ: DISCERN ölçeğine göre, her iki yapay zekâ sohbet robotu da ebeveynlerin çocukluk çağı kanserlerinde radyonüklid tedavileriyle ilgili sorularına güvenilir yanıtlar verdi. Gemini'nin yanıtları biraz daha anlaşılması daha kolay olsa da, eğitim düzeyi lise ile sınırlı yetişkinler için bu metinlerin anlaşılması zor olabilir. Ayrıca, AI-chatbotların yanıtları yalnızca genel bilgi amaçlıdır ve profesyonel tıbbi tavsiye veya klinik karar verme sürecinin yerini almamalıdır.

Anahtar Kelimeler

• Radyonüklid tedavisi
• Ebeveyn bilgisi
• Çocukluk çağı kanserleri
• ChatGPT-4
• Google Gemini Advanced
• Yapay zekâ

Etik Beyan

Mevcut çalışmada kullanılan ChatGPT-4 ve Gemini Advanced sohbet robotları genel kullanıma açık olduğundan ve bu çalışma herhangi bir hayvan veya insan katılımcı verisi içermediğinden etik kurul onayı gerekmemiştir. Literatürde de benzer bir protokol ile çalışmalar yürütülmüştür.

Are AI-Chatbots' Responses about Radionuclide Treatments for Childhood Cancers Reliable and Readable Resources for Parents?

Öz

AIM: We aimed to investigate whether two AI-chatbots could provide reliable and readable information to parents' questions about the most common radionuclide treatments in childhood cancers. MATERIAL AND METHOD: Thirty-five parental questions for each of I-131 (RAI) for thyroid cancer and I-131 MIBG for neuroblastoma were separately posed to Gemini Advanced and ChatGPT-4 in November 2024. The replies were assessed for reliability using DISCERN scale and for readability using Gunning Fog Index (GFI), Flesch Reading Ease (FRE), and Flesch-Kincaid Reading Grade Level (FKGL). The inter-rater reliability of DISCERN scores was analyzed among three reviewers: a nuclear medicine physician, ChatGPT-4, and Gemini. RESULTS: DISCERN scores (mean ± SD) rated by the physician for answers about RAI and MIBG treatments were 54 ± 6.37 and 60.8 ± 3.81 for ChatGPT-4, and 54.63 ± 6.01 and 54.77 ± 5.12 for Gemini, respectively. DISCERN scores obtained from ChatGPT-4 for answers about RAI and MIBG treatments were 50.26 ± 5.87 and 58.09 ± 4.42 for ChatGPT-4, and 59.66 ± 6.91 and 62.63 ± 4.88 for Gemini, respectively. DISCERN scores assessed by Gemini for replies about RAI and MIBG were 50.86 ± 6.33 and 54.09 ± 5.99 for ChatGPT-4, and 57.2 ± 6.97 and 56.11 ± 7.29 for Gemini. FKGL values for ChatGPT-4’s answers regarding RAI and MIBG were 13.34 ± 2.01 and 12.8 ± 2.17, respectively. FKGL values for Gemini’s responses regarding RAI and MIBG were 12.18 ± 1.47 and 11.53 ± 1.74, respectively. CONCLUSION: According to DISCERN scale, both AI-chatbots provided reliable answers to parents’ questions about radionuclide treatments in childhood cancers. The responses may be challenging for adults with only a high school education, although Gemini’s replies were slightly easier to understand. Additionally, the AI-chatbots responses are for general informational purposes only and should not replace professional medical advice or the clinical decision-making process.

Anahtar Kelimeler

• Radionuclide treatment
• Parent information
• Childhood cancers
• ChatGPT-4
• Google Gemini Advanced
• Artificial intelligence

Etik Beyan

Since ChatGPT-4 and Gemini Advanced used in the current study are open to public use and this study does not contain any animal or human participant data, the ethics committee approval was not required. In literature, studies were conducted with a similar protocol.

Kaynakça

1. 1. Moleti M, Aversa T, Crisafulli S, Trifirò G, Corica D, Pepe G, et al. Global incidence and prevalence of differentiated thyroid cancer in childhood: systematic review and meta-analysis. Front Endocrinol (Lausanne). 2023;14:1270518. doi: 10.3389/fendo. 2023.1270518. PMID: 37795368.
2. 2. Toraih E, Webster A, Pineda E, Pinion D, Baer L, Persons E, et al. Radioactive iodine ablation therapy reduces the risk of recurrent disease in pediatric differentiated thyroid carcinoma. Surg Oncol. 2024;56:102120. doi: 10.1016/j.suronc.2024.102120. PMID: 39154543.
3. 3. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2020. CA Cancer J Clin. 2020;70(1):7-30. doi: 10.3322/caac.21590. PMID: 31912902.
4. 4. Filippi L, Chiaravalloti A, Schillaci O, Cianni R, Bagni O. Theranostic approaches in nuclear medicine: current status and future prospects. Expert Rev Med Devices. 2020;17(4):331-43. doi: 10.1080/17434440.2020.1741348. PMID: 32157920.
5. 5. LaFond CM, Yost A, Lankin K, Kilaru M, Cohn SL. The Experience of Children With Neuroblastoma and Their Parents During Single-Room Isolation for (131)I-Metaiodobenzylguanidine Therapy: A Qualitative Descriptive Study. J Pediatr Hematol Oncol Nurs. 2022;39(5):304-16. doi: 10.1177/27527530211068749. PMID: 36129889.
6. 6. Pehora C, Gajaria N, Stoute M, Fracassa S, Serebale- O'Sullivan R, Matava CT. Are Parents Getting it Right? A Survey of Parents' Internet Use for Children's Health Care Information. Interact J Med Res. 2015;4(2):e12. doi: 10.2196/ijmr.3790. PMID: 26099207
7. 7. Luo X, Tong S, Fang Z, Qu Z. Frontiers: Machines vs. humans: The impact of artificial intelligence chatbot disclosure on customer purchases. Marketing Science. 2019;38(6):937-47. doi: 10.1287/mksc.2019.1192.
8. 8. Hunter RB, Thammasitboon S, Rahman SS, Fainberg N, Renuart A, Kumar S, et al. Using ChatGPT to Provide Patient-Specific Answers to Parental Questions in the PICU. Pediatrics. 2024;154(5). doi: 10.1542/peds.2024-066615. PMID: 39370900.

9. 9. Adelstein JM, Sinkler MA, Li LT, Mistovich RJ. ChatGPT Responses to Common Questions About Slipped Capital Femoral Epiphysis: A Reliable Resource for Parents? J Pediatr Orthop. 2024;44(6):353-7. doi: 10.1097/BPO.0000000000002681. PMID: 38597253.
10. 10. Cui K, Lin J, Hu Z, Hong P, Gao Z, Tian X, et al. Exploring the Potential of Large Language Models: Can ChatGPT effectively assume the role of medical professionals by providing accurate and reliable responses in childhood cancer? Research Square. 2024. doi: 10.21203/rs.3.rs-4624109/v1.
11. 11. Chen S, Kann BH, Foote MB, Aerts HJ, Savova GK, Mak RH, Bitterman DS. The utility of ChatGPT for cancer treatment information. medRxiv. 2023:2023.03.16.23287316. doi: 10.1101/2023.03.16.23287316
12. 12. Bir Yücel K, Sutcuoglu O, Yazıcı O, Ozet A, Ozdemir N. Can artificial intelligence provide accurate and reliable answers to cancer patients’ questions about cancer pain? Comparison of chatbots based on ESMO cancer pain guideline. memo - Magazine of European Medical Oncology. 2024;17(4):302-6. doi: 10.1007/ s12254-023-00954-6
13. 13. Homolak J. Opportunities and risks of ChatGPT in medicine, science, and academic publishing: a modern Promethean dilemma. Croat Med J. 2023;64(1):1-3. doi: 10.3325/cmj.2023.64.1. PMID: 36864812.
14. 14. Pan A, Musheyev D, Bockelman D, Loeb S, Kabarriti AE. Assessment of Artificial Intelligence Chatbot Responses to Top Searched Queries About Cancer. JAMA Oncol. 2023;9(10):1437-40. doi: 10.1001/jamaoncol.2023.2947. PMID: 37615960.
15. 15. Wyatt KD, Alexander N, Hills GD, Liang WH, Kadauke S, Volchenboum SL, et al. Making sense of artificial intelligence and large language models-including ChatGPT-in pediatric hematology/ oncology. Pediatr Blood Cancer. 2024;71(9):e31143. doi: 10.1002/pbc.31143. PMID: 38924670.
16. 16. Erdat EC, Köksoy EB, Utkan G. Artificial Intelligence Chatbot as a Companion for Cancer Patients About Most Common Questions: Analysis of Readability and Quality. Acta Haematologica Oncologica Turcica. 2025;58(1):21-7. doi: 10.4274/ahot.galenos. 2024.2024-8-2
17. 17. Bilgin G, Bilgin C, Childs D, Orme J, Burkett B, Packard A, et al. Performance of ChatGPT-4 and Bard Chatbots in Responding to Common Patient Questions on Prostate Cancer 177Lu-PSMA-617 Therapy. Front Oncol 2024;Jul 12;14:1386718. doi: 10.3389/ fonc.2024.1386718. PMID: 39070149.
18. 18. Aydinbelge-Dizdar N, Dizdar K. Reliability and Readability evaluation of chatbots responses as a patient information resource for the most common PET/CT scans. Rev Esp Med Nucl Imagen Mol (Engl Ed). 2024;44(1):500065. doi: 10.1016/j.remnie. 2024.500065. PMID: 39349172.
19. 19. Şan H, Bayrakcı Ö, Çağdaş B, Serdengeçti M, Alagöz E. Reliability and readability analysis of ChatGPT-4 and Google Bard as a patient information source for the most commonly applied radionuclide treatments in cancer patients. Rev Esp Med Nucl Imagen Mol (Engl Ed). 2024;43(4):500021. doi: 10.1016/j.remnie. 2024.500021. PMID: 38821410.
20. 20. Stephenson-Moe CA, Behers BJ, Gibons RM, Behers BM, Jesus Herrera L, Anneaud D, et al. Assessing the quality and readability of patient education materials on chemotherapy cardiotoxicity from artificial intelligence chatbots: An observational cross-sectional study. Medicine (Baltimore). 2025;104(15):e42135. doi: 10.1097/MD.0000000000042135. PMID: 40228277.
21. 21. Sharma S, Latif Z, Makuvire TT, Taylor CN, Vargas F, Abo-Sido NS, et al. Readability and Accessibility of Patient-Education Materials for Heart Failure in the United States. J Card Fail. 2025;31(1):154-7. doi: 10.1016/j.cardfail.2024.06.015. PMID: 39094729.
22. 22. Matic R, Kabiljo M, Zivkovic M, Cabarkapa M. Extensible Chatbot Architecture Using Metamodels of Natural Language Understanding. Electronics. 2021;10(18):2300. doi: 10.3390/electronics10182300.
23. 23. Drapkin A. Google Gemini vs ChatGPT 2024: AI Chatbot Head-to-Head Test 2024 [updated March 13, 2024. Available from: URL: https://tech.co/news/google-bard-vs-chatgpt. (Accessed October 10, 2025).
24. 24. Rogasch JMM, Metzger G, Preisler M, Galler M, Thiele F, Brenner W, et al. ChatGPT: Can You Prepare My Patients for [(18)F]FDG PET/CT and Explain My Reports? J Nucl Med. 2023;64(12):1876-9. doi: 10.2967/jnumed.123.266114. PMID: 37709536.
25. 25. Durmaz Engin C, Karatas E, Ozturk T. Exploring the Role of ChatGPT-4, BingAI, and Gemini as Virtual Consultants to Educate Families about Retinopathy of Prematurity. Children (Basel). 2024;11(6). doi: 10.3390/children11060750. PMID: 38929329.
26. 26. Patil NS, Huang R, Mihalache A, Kisilevsky E, Kwok J, Popovic MM, et al. The ability of artificial intelligence chatbots Chat- GPT and Google Bard to accurately convey pre-operative information for patients undergoing ophthalmological surgeries. Retina. 2024 44(6):950-3. doi: 10.1097/IAE.0000000000004044. PMID: 38215455.
27. 27. Nash E, Bickerstaff M, Chetwynd AJ, Hawcutt DB, Oni 63 L. The readability of parent information leaflets in paediatric studies. Pediatr Res. 2023;94(3):1166-71. doi: 10.1038/s41390-023- 02608-z. PMID: 37120650.