Minimally Invasive versus Abdominal Radical Hysterectomy for Cervical Cancer
2018 · paper · Clinical trials
A randomized trial comparing minimally invasive with open radical hysterectomy in 631 women with early-stage cervical cancer. Disease-free survival at 4.5 years was lower with minimally invasive surgery (86.0% vs 96.5%), as was overall survival.
2 findings
Wording Text gives the combination of adjuvant chemotherapy and radiation as "18.8% in the minimally invasive surgery group and 18.1% in the open-surgery group", but Table 2 counts imply 53/312 (17.0%) and 61/319 (19.1%), and 18.1% is not any whole number out of 312.
- p. 5 Results, adjuvant therapy paragraph
chemotherapy and radiation (18.8% in the minimally invasive surgery group and 18.1% in the open-surgery group)
- p. 5 Table 2, open surgery column
86 (27.6)
- p. 5 Table 2, open surgery column, chemotherapy row
66 (21.2)
- p. 5 Table 2, open surgery column, radiotherapy row
73 (23.4)
Only the authors can say what is right here.
Judge's note: Read Table 2 and the paragraph on the page 5 image: open 86 either, 66 chemotherapy, 73 radiotherapy; minimally invasive 92, 72, 81. Inclusion-exclusion gives 53 and 61 patients with both. 56/312 is 17.9% and 57/312 is 18.3%, so 18.1% cannot come from N=312; 18.8% matches 60/319, one fewer than Table 2 implies. No definition in the article explains the gap. The text cites Table S3, which is not among the files, and may use a different denominator (for example 51/282 = 18.1% among open-arm patients who had surgery) or a narrower definition such as concurrent chemoradiation. The mismatch between the text and Table 2 is real as printed; only the authors or Table S3 can say which basis applies. Differences are about one percentage point and do not change the conclusion of similar adjuvant therapy.
Second check: another model agreed
Report a problem with this finding
Wording Discussion: "at the time of analysis, 59.7% of the patients had reached the 4.5-year time point (median follow-up, 2.5 years)"; the Results define 59.7% as "the rate of available survival information on the primary outcome at 4.5 years", and a 2.5-year median follow-up with 90 + 80 at risk at 4.5 years rules out 59.7% of patients having reached 4.5 years.
- p. 9 Discussion, limitations paragraph
59.7% of the patients had reached the 4.5-year time point (median follow-up, 2.5 years)
- p. 5 Results, Survival Outcomes, first sentence
the rate of available survival information on the primary outcome at 4.5 years was 59.7%
- p. 6 Figure 1B, No. at Risk, 4.5 years
90
Judge's note: Verified the Discussion sentence on the page 9 image, the Results sentence on page 5 and the Figure 1B numbers at risk (90 and 80 at 4.5 years) on page 6. 170 of 631 is 26.9%; adding the 34 patients with a disease-free survival event gives 32.3%. A median follow-up of 2.5 years means at most half the patients were followed beyond 2.5 years. The 59.7% is an information (data-maturity) fraction per reference 25, as the Results say; the Discussion misdescribes it as a share of patients. The fix is to use the Results wording.
Second check: another model agreed
Report a problem with this finding
3 minor (typos and the like; most checked by the judge only)
Reference or label Figure 2 is titled "Kaplan–Meier Estimates of Overall Survival, Disease-Specific Survival, and Locoregional Recurrence", but its legend says panels B and C are "cumulative incidence curves" from a Fine and Gray competing-risks model; panel B is headed "Disease-Specific Survival" while plotting cumulative incidence of death from cervical cancer.
- p. 8 Figure 2 title
Kaplan–Meier Estimates of Overall Survival, Disease-Specific Survival, and Locoregional Recurrence.
- p. 8 Figure 2 legend, panel B
Panel B shows the cumulative incidence curves for dis-ease-specific survival
Judge's note: Viewed Figure 2 on the page 8 image. Panel A is a Kaplan–Meier plot (proportion alive). Panels B and C have y-axes "Cumulative Incidence of Death from Cervical Cancer (%)" and "Cumulative Incidence of Locoregional Recurrence (%)", and the legend describes them as competing-risks cumulative incidence curves. The title is therefore accurate for panel A only. The data and hazard ratios are unaffected; a loose figure title.
Second check: another model agreed
Report a problem with this finding
Wording Abstract lists the adjustment covariates as "age, body-mass index, stage of disease, lymphovascular invasion, and lymph-node involvement", omitting the ECOG performance-status score that the Results text (p5) and Table 3 footnote (p7) include.
- p. 1 Abstract, Results
adjustment for age, body-mass index, stage of disease, lympho-vascular invasion, and lymph-node involvement
- p. 7 Table 3, footnote *
lymphovascular invasion, lymph-node involve-ment, and ECOG performance-status score
- p. 7 Table 3 footnote *
lymph-node involve-ment, and ECOG performance-status score.
Judge's note: Checked the abstract on the page 1 image, the Results paragraph on page 5 and the Table 3 footnote on page 7. The Results and footnote both give six covariates ending in "and ECOG performance-status score"; the abstract gives five as a closed list. The result is not affected; the abstract's list is incomplete.
Second check: another model agreed
Report a problem with this finding
Wording 'The lower boundary of the confidence interval included the noninferiority margin of −7.2 percentage points' is loose wording: a boundary (−16.4) cannot include a value. The interval (−16.4 to −4.7) contained the margin, and its lower boundary lay below it.
- p. 5 Results, Survival Outcomes, right column
The lower boundary of the confidence interval included the noninferiority margin of −7.2 percentage points
- p. 5
The lower boundary of the confidence interval included the noninferiority margin
Judge's note: Checked p5 right column. The printed CI is −16.4 to −4.7 and the margin is −7.2. The abstract (line 33) and the Fig. 1 legend give the criterion as the lower boundary lying above the margin. Here the lower boundary falls below the margin and the interval spans it, so the conclusion ('noninferiority was not declared') is correct. Only the phrasing is imprecise, and the meaning can be recovered.
Second check: another model agreed