Castle Biosciences, Inc. — Q2 2019 Form 10-Q
Reporting period: Three and six months ended June 30, 2019; interim financial statements are unaudited. Castle is a commercial-stage dermatological cancer diagnostics company, primarily selling DecisionDx-Melanoma and DecisionDx-UM.
Financial performance
| Metric | Q2 2019 | Q2 2018 | Six months 2019 | Six months 2018 |
|---|---|---|---|---|
| Revenue | $10.74 million | $3.98 million | $19.46 million | $7.64 million |
| Gross margin | $8.75 million; 81.4% | $2.65 million; 66.7% | $15.86 million; 81.5% | $5.06 million; 66.2% |
| Operating income (loss) | $0.61 million | $(2.67) million | $0.29 million | $(5.75) million |
| Net loss | $(1.27) million | $(3.15) million | $(2.63) million | $(6.79) million |
| Net loss attributable to common stockholders | $(2.26) million | $(4.08) million | $(4.61) million | $(8.57) million |
- Revenue rose 169.9% in Q2 and 154.7% for the first half, reflecting increased test volume, higher per-test revenue and the impact of Medicare coverage and reimbursement for DecisionDx-Melanoma. Reported revenue also included prior-period variable-consideration adjustments: positive $3.3 million in Q2 2019 and $2.8 million in the first half.
- Test reports delivered increased 23.9% year over year in Q2 and 21.0% in the first half. First-half reports totaled 7,659, versus 6,330 a year earlier.
- SG&A increased 64.0% in Q2 and 53.4% in the first half, partly due to sales-force expansion from 14 to 23 territories. Interest expense increased significantly, including noncash discount amortization on the Q1 convertible notes.
- Cash from operations was $1.77 million in the first half, compared with cash used of $7.21 million in 2018. The 2019 inflow was supported by a $3.63 million decrease in accounts receivable and noncash adjustments; it does not represent net income.
Balance sheet, debt and liquidity
- At June 30, cash and cash equivalents were $17.47 million; current assets were $28.52 million and current liabilities $11.75 million. Total assets were $33.35 million. Accumulated deficit was $60.12 million.
- Long-term debt, net of discounts, was $24.69 million, including $0.83 million current. The company also reported $4.43 million of current convertible promissory notes, including an embedded derivative; the notes had $11.77 million principal outstanding. Substantially all assets, except intellectual property and specified exceptions, secure the term loan.
- In June, the company amended its loan agreement: it eliminated the $5 million revolving line, increased term borrowing by $5 million, and changed the monthly revenue covenant from a trailing six-month to a trailing three-month test. The company reported compliance and management expected compliance over the next 12 months. Term debt carried an 8.88% floating rate at quarter-end, with principal repayments scheduled to begin June 2020 and maturity in November 2022.
- After quarter-end, the July 29 IPO raised approximately $66 million net from 4.6 million shares sold at $16 per share. The Q1 notes and a subsequent $10 million July note converted into common stock at the IPO; preferred stock and warrants also converted. The company stated that IPO proceeds, existing cash and expected test-sale collections should fund operations for at least 24 months.
Outlook, risks and notable items
- No formal revenue or earnings guidance is provided. Management expected significant expenses and operating losses for at least the next two years, and planned to launch two late-stage product candidates in the second half of 2020, subject to development and commercialization risks.
- DecisionDx-Melanoma and DecisionDx-UM remain reliant on third-party reimbursement. In the first half, Medicare accounted for 45% of revenue, Medicare Advantage plans 22%, UnitedHealthcare 9% and BlueCross BlueShield plans 7%. Coverage decisions, collections and revenue estimates may cause substantial period-to-period variability.
- A consolidated appeal involving 2,698 DecisionDx-Melanoma Medicare claims was heard in April and May 2019; no formal ruling had been issued. Revenue for the claims remained fully constrained pending the administrative law judge’s decision.
- CMS granted DecisionDx-Melanoma new ADLT status: the Medicare rate was set at the $7,193 list price from July 2019 through March 2020, with later rates based on private-payor data. CMS may recoup amounts if the list charge exceeds 130% of the weighted median private-payor rate. DecisionDx-UM received existing ADLT status, with its Medicare rate set by Noridian through 2020 and based on private-payor rates beginning in 2021.
- The 2017 and 2018 audit reports included a going-concern explanatory paragraph because of covenant-compliance concerns. Following the June loan amendment, management said it expected covenant compliance; the filing still describes substantial historical losses and financing risks.
- Disclosure controls were not effective as of June 30 due to material weaknesses involving journal-entry review, accounting expertise and ASC 606 variable-consideration controls. Remediation efforts were underway, but completion was not established.
- Other key risks include changes in LDT regulation, reliance on a single laboratory and sample database, product development and adoption uncertainty, and variable-rate debt exposure. The company estimated a 1 percentage-point rate increase would raise annual interest expense and reduce operating cash flow by approximately $250,000.
Important facts for investors to verify
- How much of reported revenue growth reflects recurring test volume and reimbursement versus prior-period variable-consideration adjustments?
- What was the outcome and financial impact of the consolidated Medicare claims appeal, and are revenue-covenant requirements being met?
- How are ADLT reimbursement rates, private-payor coverage and collections developing, including any potential CMS recoupment?
- Are the IPO proceeds and stated 24-month runway sufficient given planned commercial, R&D and public-company spending?
- Have the material weaknesses in financial reporting controls been remediated, and are reported results reliable as estimates and processes evolve?