A total of 57 patients were included in the data analysis. Patients’ ages ranged from 2 to 21 years of age at time of cancer diagnosis. Twenty-nine patients were less than 13 years of age at time of cancer diagnosis and the remaining 28 patients were 13 years or older. For patients less than 13 years of age, 17 were found to have leukemia diagnoses (T or B cell) and 12 were found to have either Hodgkin lymphoma or a solid tumor diagnosis. For patients 13 years or older, 4 were found to have been diagnosed with leukemia and the remaining 24 had either a Hodgkin lymphoma or solid tumor diagnosis. Solid tumor diagnoses included Wilms tumor, Ewing’s sarcoma, osteosarcoma, primary mediastinal large B cell lymphoma, germ cell tumor of the ovary, hepatoblastoma and undifferentiated sarcoma.
Premature ovarian insufficiency was identified in 5 patients. All 5 patients were found to be 13 years old or older. A Fisher’s exact test indicated statistical significance (p = 0.012) when comparing the two age groups and the presence of POI. Although all 5 patients with POI fell into the lymphoma/solid tumor category, no statistical significance (p = 0.15) was found when comparing cancer type (leukemia vs lymphoma/solid tumor) and presence of POI. This is likely attributable to the small patient population in each subgroup.
To evaluate associations between CED, age and presence of POI, a scatter plot was created (Fig. 2). As having POI was binary (yes or no) and CED was found to be not normally distributed, a nonparametric Spearman correlation was conducted to assess these variables with the primary outcome of interest. The correlation coefficient was not found to be relatively strong at R = 0.425 for CED vs age, however the P value was significant at 0.001 indicating that the higher CED values were associated with older ages. The majority of the lymphoma/solid tumor patients in our analysis were 13 years & older (n = 24). When assessing CED with POI, we did not find a statistically significant result (R = 0.225, p = 0.096). Interestingly, of those with POI, the CED ranged from 0 to 28.4gm/m2. As similar to the previous analysis, there was an association between older age and POI (R = 0.317, p = 0.017).
As the values for the continuous variable were quite broad, Table 1 shows the statistical summary for each hormone value and CED. Nine patients had AMH values available to report. These were not used in any analysis given the limited number. Seven out of the 9 patients were seen by REI and had levels drawn in their fertility evaluation. The range of AMH values was < 0.03–5.07 ng/mL. Of these 9 patients, only 1 patient was deemed to have POI given the levels of the other hormone values. This patient’s AMH was found to be < 0.03. A second patient’s AMH value was found to be 0.74, however, they were excluded from the original analysis due to only having FSH documented. Conversely, a third patient’s AMH value was reported at 0.7, however the remainder of their hormone evaluation and menses history was within normal limits.
Pregnancy was reported when information was made available. Twelve patients had documented live births. Of these 12, one patient was referred to REI and found to have an AMH of 0.74 with a history of abnormal periods. All 12 patients had received their diagnosis and chemotherapy when they were 13 years or older.
Menses history was recorded when found in documentation within the electronic record. Thirty patients separate from those used in the ovarian hormone analysis had menses history available following completion of chemotherapy. Due to the limited patient number, analyses were unable to be conducted. Twelve of the 30 patients were 13 years or older. Of the 30, only 5 patients had reported irregular menses. CED totals were available for 4 out of the 5 patients with a range of 0.33–8.3 gm/m2.