一份檔案,三件事:上面是這三科的期末排程,中間是 102 的完整架構(含 L7–L9),最下面是你的學習協定、考後紅線卡,以及這套東西怎麼接到 BD / MSL 的職涯上。所有內容 inline,離線也能開。
English answers are written out under every Chinese line because the exam is written in English.
七份 deliverable、兩個戰區、52 天。下面每一個日子都算過是星期幾,不是估的。你的可用時數設定:平日 2–4 小時、週末整天、週五 18:30–21:30 要上課。
你貼的時間表說明一件事——101 和 102 都還沒結束。
| 科目 | Exam | Date | Time | Venue | D- |
|---|---|---|---|---|---|
| ABCT5101 Modern Approaches in Biopharmaceutical Development | Paper 01 | 2026-12-09(三) | 19:00–21:00 | FSCH | 44 |
| ABCT5102 Pharmacology and Toxicology in Biotherapeutics | Paper 01 | 2026-12-17(四) | 19:00–21:00 | SH2 | 52 |
三個推論:
FSCH / SH2 是場地代號,開考前兩週上 eStudent 確認一次就好,現在不用管。
103 的論文(40%)和演示 PPT 同樣在 11 月 9 日(星期一)截止——距今 14 天。12 月那兩場考試不會幫你分擔這 14 天。
而且 11/9 是星期一,意思是最後一個完整週末是 11/7–11/8,交件前一天沒有週末可以救。這條決定了第一戰區的全部排程。
| 區 | Course | Deliverable | Weight | Due | 你能控制 |
|---|---|---|---|---|---|
| 1 | 101 | MC quiz | 待確認 | 本週 10/26–11/1 | 100% |
| 1 | 103 | Individual essay · 1,000 words | 40% | 11/9 Mon | 100% |
| 1 | 103 | 你個人的 4–5 分鐘 PPT segment | ≈35%(50%×70%) | 11/9 Mon | 100% |
| 1 | 102 | Presentation | 待確認 | 11/9 前 | 100% |
| 1 | 103 | 小組協調(Group 8 · 13 人) | 15%(50%×30%) | 11/9 PPT | 約 8% |
| 2 | 101 | Exam Paper 01(2 小時) | 待確認 | 12/9 Wed | 100% |
| 2 | 102 | Exam Paper 01(2 小時) | 待確認 | 12/17 Thu | 100% |
排優先序只有一條規則:權重 × 你的控制權。所以 103 論文排第一(權重最高、完全自己掌控),小組協調排最後(權重最低、你幾乎管不到)——但它必須最早開口,因為它會反過來堵死你。
兩個戰區中間有 11/21–12/8 整整 19 天空檔。那是這份計畫最容易崩掉的地方——不是因為忙,是因為不忙。
演示題目 18 條、論文題目 134 條,是兩份不同清單,但你可以選同一個主題,用同一批文獻交兩份功課。文獻只查一次,省約 8 小時。
推薦這條(v.3 演示清單第 17 條):
"Dissecting the Deal: How IP Strength Dictates Licensing Terms in Biotech Partnerships"理由很直接:你每天在寫 BD Daily 的 licensing deals,這題是你已經在做的工作,不是新知識。而且 Q&A 時你講的是自己的地盤,Engagement 那 10% 基本白拿。
備案兩條:"Monetizing a Patent Portfolio: Licensing, Cross-Licensing, and Sales"(同路線,偏財務)/"Platform vs. Product: Tailoring IP Strategy to the Business Model"(對接你想去的儀器商)。
目標:101 考完,103 題目定死,文獻到手,Group 8 開口。
| Day | Block | 做什麼 |
|---|---|---|
| Mon 10/26 今晚 | 2h + 15min | ① 101 第一輪(直接用 drill-abct101.html,MC 是辨認不是回想,對你這種模式有利)② 15 分鐘在 Group 8 群組發話:我佔 [題目],誰來統整分工表——今晚就發,不要等 |
| Tue 10/27 | 2h + 1h | 101 第二輪 + 把 103 題目鎖死 |
| Wed 10/28 | 2h | 101 最後一輪,計時 mock;考完就放手,不再回頭 |
| Thu 10/29 | 2h | 103 文獻第一批:目標 6–8 篇(2 篇 review + 3–4 篇案例 + 1 篇條文原文) |
| Fri 10/30 | — | 18:30–21:30 上課(103 L8 · PCT),今晚不算讀書時間 |
| Sat 10/31 | 6h(切 3 塊) | 文獻讀完 + 論文 outline 寫死:claim → evidence → implication,每段一句 |
| Sun 11/1 | 4h | 引言 + 第一個主體段落初稿;順手把 102 presentation 的規格問清楚 |
✅ Checkpoint W1:101 已考 / 103 題目鎖死 / 8 篇文獻到手 / outline 完成 / Group 8 已開口 / 102 格式已知
Checkpoint 沒過就不要往下走,下一週的排程是建立在這個基礎上的。
目標:論文定稿、週日就交、你的 segment 成型。
✅ Checkpoint W2:論文 + PPT 已送出(有截圖或收據)/ 你的簡報 6–8 頁成型 / 102 完成
⚠️ 這週需求量約 22h、可用約 24h,緩衝只有 2 小時。如果炸掉:砍 102 的版面,不要砍內容,也不要砍 103。
✅ Checkpoint W3:兩份已送出 / 練講 3 次 / 計時 ≤ 5 分鐘(超時就砍段落,不要加快語速)
✅ Checkpoint W4:103 演示完成 / 第一戰區全部清空
這一週的任務是「不要開始新東西」。你 11/20 晚上剛講完,接下來是時間表上第一段沒有死線的日子——而這正是最容易崩的地方。
✅ Checkpoint W5:101 六個 lecture 各跑過一次 drill / 手上有一張「弱點清單」,不超過 6 條
✅ Checkpoint W6:計時 mock 達到你的目標分 / 弱點剩 ≤ 3 條 / 知道哪幾章可以直接放棄
✅ Checkpoint W7:101 考完,然後當晚不要覆盤,直接睡。你還有 8 天。
✅ Done:三科全部結清。之後時間表上沒有你的名字。
兩場都是 19:00–21:00。意思是:上班一整天 → 通勤 → 直接進考場,中間沒有真正吃飯的窗口。這正是 10/22 讓你幾乎昏倒的同一個設定——空腹 + 高壓 + 兩小時久坐。時間表把這個陷阱複製了兩次。
三條硬規則,不談判:
你的知識量從來不是問題,10/22 已經證明過了。12 月這兩場,贏在生理,不是贏在筆記。
| Item | Hours | 排在哪 |
|---|---|---|
| 101 MC quiz | 8–10 | W1 |
| 102 presentation | 8–12 | W1 末 – W2 |
| 103 個人論文 | 18–20 | W1 Sat – W2 Sun |
| 你的 PPT segment | 5–6 | W2 Sun – W3 |
| Group 8 協調 | 2–3 | W1 起,持續但低強度 |
| 練講(含彩排) | 4 | W3 – W4 |
| 101 期末考準備 | 18–22 | W5 – W7 |
| 102 期末考準備 | 14–18 | W8 |
名目總容量遠大於這個數字,所以瓶頸不是時間——是啟動成本、是被 13 人卡住、以及 12 月那段 19 天的空檔。下面六條就是在處理這三件事。
這四條的作用是同一件事:把「想起來我在幹嘛」的成本從每次都付,降成一天付一次。
最後一條是你的單一最大風險:Group 8 有 13 個人,而你只有 8% 的控制權。W1 週一就要開口,不是等到 W2。
第一項會改變 W8 的分配,第二、三項會改變 W1–W2 的比例,其餘只影響風險。但目前這份表已經可以照著跑。
Group 8 的 PPT 題目已經定案,所以原本那三條「協調型」目標作廢。新設定只有一句:不去動別人的決定。精力全部放在你 100% 掌控的兩件事——你自己的 1,000 字,和你講的那 4–5 分鐘。
副作用是好的:13 個人寫同一個題目,代表你的論文不可能靠「選到好題目」勝出,只能靠角度——而你的角度是 L7–L9,其他人不會有。
S-103a(Individual essay 1,000w · 40% · 11/9)abct103-essay92.html。S-103a ⇄ S-102b(12/17 Paper 01)真正在做的事:把「為了論文讀的東西」轉成「為了考試存的 index 句」。這是你 CPU 強、RAM 小最該用的一次。
S-103b(你的 4–5 分鐘 · deck=Task3_Humira_Intro_Slides_EN.pptx)· S-103c(Group 8,你控制率約 8%)S-101b(12/9 FSCH)· S-102b(12/17 SH2)順帶:原本的「把作業變成職涯資產」降為 G5′ 第五段的副產物——Evergreening 這題本身就是一篇可發的 BD Lens,不用另外起題。
這是兩個不同題目,不要再混在一起:論文=Topic #92(NGS 診斷的 IP),PPT=Group 8 的 Evergreening(Humira 個案)。論文已寫滿 1,065 字,本體在下面那個檔案裡,不在這一頁。
→ abct103-essay92.html 8 段 · 1,065 字(上限 1,100)· 貼進 Word 用 Calibri 12 / 雙行距再數一次,引用不計。
題目:The Role of IP in Diagnostics and Medical Devices Development(Topic #92)
Thesis:定價權不在分析性能,而在 claims 怎麼分佈在三層——instrument → reagent/chemistry → algorithm & interpretation;而你那層能不能變成可排他的資產,取決於 RUO → IVD 的法規路徑,以及各法域怎麼對待診斷方法 claims。
| ¶ | 段 | 這一段在做什麼 | 字 |
|---|---|---|---|
| 1 | Intro | 把「一台機器」拆成三層可分別主張的保護 | 127 |
| 2 | Intro | Thesis + 四個法域(US / EU / 中國內地 / HK) | 125 |
| 3 | Layer A+B | 市場證據:2019 年 41.38 億美元、Illumina 74.1% | 127 |
| 4 | Layer A+B | 服務型 lab 的處境:三層裡只真正生產 Layer C | 136 |
| 5 | HK 判例 | Illumina v MGI(2020-08):跑服務本身可能就是侵權行為 | 145 |
| 6 | 管轄 | HK 標準專利=再註冊;HK 不是 PCT 成員 | 129 |
| 7 | 管轄 | RUO→IVD + Ariosa v Sequenom / EPC Art. 53(c)、G 1/04 | 137 |
| 8 | 結論 | 合成 + 一句前瞻(你要站在哪一層) | 139 |
| Total | 1,065 | ||
可接受範圍 900–1,100。上面是實測,不是預算。
組裡的 deck 是 Task3_Humira_Intro_Slides_EN.pptx,四部分:① What Humira is(adalimumab、TNF-α、2002-12 首次批准、9 個後續適應症、40 mg/2 週、$522→$1,024→$2,984)② How the price became what it is(27 次加價、淨價 +110%、shadow pricing)③ The orphan-drug route(8 個孤兒藥指定、7 年獨占、salami slicing)④ So is this innovation, or an anti-competitive strategy?(判準頁:同一個分子可以是兩者,分界在 increment)。①②③ 的事實出自 US House Oversight 2021 報告、頁碼 deck 上已標;④ 的判準句是論證,不是引用。
你的角度(不碰別人的題目):接 Part 2 → Part 3 的轉折,用一句話把三段串起來——同一個分子,靠的不是新科學,是把保護的形狀重新安排一次。加價 27 次與 8 個孤兒藥指定,是同一個動作的兩個價錢。
≤5 張 slide、≤450 字口說稿、11/20 前練 3 次。只交付你的 segment,不改組員的結構與題目。
論文練的是「誰擁有 claim」,PPT 練的是「同一個分子被重新包裝幾次」——同一件事的兩面。12/17 考 L7–L9 的時候,你手上已經有一個現成的生物藥個案(Humira:生物藥、皮下注射、formulation 改變)與一個現成的安全論述,比別人多一層。這是 G2′ 現在真正的槓桿。
九堂課不是九個袋子。把它們切成五段之後,每一段都同時是生命科學的一段、藥學的一段、生物科技的一段、商業的一段——而且順序是單向的,走完第五段就到了資產負債表。
你真正要記的東西很少:七層的順序不變,只有第 3、4 層換名字。下面五段只是把「第 3、4 層換成誰」列出來而已。所以九堂課的重擔不在內容,在知道自己在第幾段。
濃度曲線 → 接收器引擎 → 最難的靶 → 型態與風險 → 證據與歸屬
同一條:It exists → It behaves → It is safe enough → It is defensible → It is worth owning。這就是你那篇 LinkedIn 的主軸,也是 ¶8 的收尾。
狀態:L1–L6 的期末申論已在 2026-10-22 考完。L7–L9 不在那次考試範圍,是下一章。下面的 L7–L9 已經用同一套七層框架寫好了。
一句話:這一科最難的那關(當場寫四題申論)你已經過了。剩下的是準備型評分——正好是你「CPU 強、RAM 小」該贏的題型。
你當天空腹應考、考後近乎昏倒。這不是心理素質問題,是低血糖加上交感風暴之後的副交感反彈。你教科書上的內容在你自己的身體上演了一遍——而且順帶提醒你一件事:大神經裡踩油門的是 NE 和 Epi,ACh 是煞車。你說「ACh 一直踩交感油門」,方向記反了,剛好是你最熟的那個框架。
下次考試前 60–90 分鐘吃碳水+蛋白。這件事的效果大於多背二十張卡。
這六條不是通用讀書建議,是從你這段對話裡的行為逐條萃取出來的。你每次卡住的原因都在這裡。
你卡住最久的地方是「blocker 到底是加速還是減速」。答案是它取決於切的是油門還是煞車。踩油門 ≡ 切煞車,切油門 ≡ 踩煞車——blocker 這個字本身不含方向資訊。
Never classify a drug as stimulant or depressant. Ask which pedal it touches: pressing the gas equals cutting the brake, and cutting the gas equals pressing the brake.
你自述 CPU 強、RAM 小。所以每一層只存一句中文因果句+一句英文對應句,細節查得回來就行,不要整段吞。
Store one causal sentence per layer, not the full story. Detail lives in the backend.
Q1 你寫出稀釋卻沒寫到排泄;Q3 你寫出受體卻沒落到血壓。機制寫完一定要再走一步到臨床結果,那一句就是分數。
After writing the mechanism always take one more step to the clinical outcome. That last step is where the marks sit.
教授明說不考藥名。寫 Class IB sodium channel blocker 比寫一個拼錯的藥名安全得多,而且還能證明你懂機制。
When the name is uncertain, write the class plus the mechanism. It is safer and it proves understanding.
Q2 的 physiological 和 Q4 的 arrhythmia,你都不是不會,是沒停下來確認問什麼。落筆前花五秒圈出題目的類別關鍵字。
Spend five seconds circling the category keyword before writing. Two of your four losses were category errors, not knowledge gaps.
你當天空腹應考,交感風暴把 RAM 佔滿,發生的是 retrieval failure 不是失憶。考前 60–90 分鐘吃碳水+蛋白,這件事的效果大於多背二十張卡。
Physiology is part of working memory. Fasting plus sympathetic storm causes retrieval failure, not memory loss. Eat carbs and protein 60 to 90 minutes before an exam.
每個器官都有兩個踏板:交感是油門,副交感是煞車。藥物只做兩件事中的一件,而且只碰其中一個踏板——踩它或切它。
| Drug hits a SYMPATHETIC receptor | Drug hits a PARASYMPATHETIC receptor | |
|---|---|---|
| Agonist(踩) | sympathetic effect | parasympathetic effect |
| Antagonist(切) | parasympathetic effect | sympathetic effect |
A blocker is not a downer.「blocker」這個字本身不含方向資訊。它只告訴你「切」,切完是加速還是減速,取決於你切的是哪個踏板。
α1 vs α2 一句話:α1 在效應器肌肉上,是「訊號落在哪裡」;α2 在神經末梢上,是「訊號送出多少」。α1 是踏板,α2 是調速器。
| Organ | Gas(交感) | Brake(副交感) | Cut the gas | Cut the brake |
|---|---|---|---|---|
| Heart | β1 → ↑rate, ↑force | M2 → ↓rate | slows | speeds up |
| Bronchi | β2 → dilate | M3 → constrict | constricts | dilates |
| Blood vessels | α1 → constrict | none | dilates | — |
| Gut | α/β → slow | M3 → contract | speeds up | slows |
| Bladder | β → relax wall | M3 → empty | empties | retains |
| Pupil | α1 → dilate | M3 → constrict | constricts | dilates |
| Sweat gland | ACh on M3(唯一例外) | — | — | — |
為什麼這張表比藥單有用:non-selective β blocker 切掉肺部 β2 的油門,那裡有煞車但沒人踩,所以氣管收縮——這就是 asthma 的禁忌。它從表裡推出來,不需要藥名。
順序是唯一要記死的東西:是什麼 → 怎麼走 → 綁什麼 → 然後呢 → 哪裡相反 → 藥物怎麼介入 → 阻斷的代價。換系統時順序不變,只有第 3、4 層換名字。
PK 回答的是「身體對藥物做了什麼」,與 PD(藥物對身體做了什麼)相反。藥物從進入到離開身體會經過四個階段:吸收、分佈、代謝、排泄。 Pharmacokinetics describes what the body does to the drug, and passes through four stages: absorption, distribution, metabolism, excretion.
口服藥物走完整條路:腸道吸收 → 門靜脈 → 肝臟 → 全身循環 → 腎臟或膽汁。靜脈注射跳過前段,所以是 100% 基準。口服藥會先經過腸壁與肝臟的首次通過代謝,因此損失生物利用度。 An oral drug travels gut to portal vein to liver to circulation to kidney or bile. Intravenous dosing skips the early steps, which is why it is the 100% baseline, while oral drugs lose bioavailability to first-pass metabolism in the gut wall and liver.
藥物被血漿蛋白結合限制。結合態分子體積大、無法通過毛細血管壁,所以只有游離型才有活性、能通過血腦屏障、並被腎小球濾過。 Protein binding limits the drug. The bound form is large in volume and cannot cross capillary walls, so only the free fraction is active, crosses the blood-brain barrier, and is filtered by the glomerulus.
代謝分兩階段。Phase I 由 CYP450 主導,多數不改變活性,但藥物之間的交互作用主要發生在這裡,因為酵素的誘導或抑制會改變藥物濃度。Phase II 做結合反應,把產物變成水溶性以便排泄。 Metabolism has two phases. Phase I is driven by CYP450 and usually preserves activity, but it is where most drug interactions occur because enzyme induction or inhibition changes drug levels. Phase II conjugates the product so it becomes water-soluble for excretion.
弱酸與弱鹼的處理方向完全相反。弱酸性藥物在酸性尿液中未解離、脂溶性高,所以被再吸收回血液,因此鹼化尿液可加速排出。弱鹼性藥物則相反,需要酸化尿液。 Weak acids and weak bases are handled in opposite directions. A weak acid is unionised in acidic urine and lipid-soluble, so it is reabsorbed, which means alkalinising the urine speeds its removal. Weak bases need the opposite treatment.
給藥途徑的選擇就是介入方式。舌下給藥繞過首渡,因為首渡代謝會在藥物到達循環前就把它破壞,所以不能吞服的藥物改用舌下。藥物設計也可以介入,例如去兒茶酚環的改造讓藥物逃過酵素降解。 Route of administration is the intervention. Sublingual dosing bypasses first-pass metabolism, which would otherwise destroy the drug before it reached circulation. Structural design also intervenes, as removing the catechol ring lets a drug escape enzymatic degradation.
分子量決定了唯一的出口,所以不是每種藥都能走腎臟。約 300 Da 以下的藥物走尿液,超過 500 Da 走膽汁,因此中間分子量的藥物需要代謝轉換才能排泄。另外與 Probenecid 併用會抑制主動分泌,讓 Penicillin 濃度升高。 Molecular weight decides the only available exit. Drugs under about 300 daltons leave through urine and those over 500 through bile, so mid-range drugs must be metabolised first. Co-administration with probenecid also inhibits active secretion and raises penicillin levels.
藥效學是藥物與受體結合後產生效應的系統,核心是一個可逆的化學平衡。因為平衡可逆,所以提高濃度可以把受體全部佔據,阻斷劑也只能把效應推回來。 Pharmacodynamics is the system that converts drug-receptor binding into effect, built on a reversible equilibrium. Because it is reversible, raising concentration can occupy every receptor, and an antagonist can only push the effect back.
基本式是游離藥物濃度加上游離受體,結合成複合體後產生效應。游離濃度而非總濃度才是有效濃度,所以藥物結合蛋白後實際能作用的量會下降。 The core equation is free drug plus free receptor forming a complex that produces effect. Free rather than total concentration is what matters, so protein binding reduces the amount actually available.
受體分四大類。離子通道型(菸鹼型 N)最直接;GPCR 分三種耦合;核受體在細胞內、作用慢但持久;受體酪胺酸激酶如胰島素。 Receptors fall into four families. Ion channels such as the nicotinic receptor are the most direct; GPCRs come in three couplings; nuclear receptors sit inside the cell and act slowly but durably; receptor tyrosine kinases such as insulin complete the list.
G 蛋白的亞型決定了第二傳訊者,而第二傳訊者決定方向。Gq 走 IP₃ 與 DAG 並升高鈣離子;Gi 降低 cAMP;Gs 升高 cAMP。口訣是 QISS。 The G-protein subtype determines the second messenger, and the second messenger determines direction. Gq raises calcium through IP₃ and DAG, Gi lowers cAMP, and Gs raises cAMP.
Kd 不等於 ED₅₀,因為一個描述結合、一個描述效應,而效應還受備用受體與訊號放大影響。效價與效能也互相獨立,所以效價高不代表效能高。 Kd does not equal ED₅₀ because one describes binding and the other describes effect, and the effect also depends on receptor reserve and signal amplification. Potency and efficacy are likewise independent.
激動劑與拮抗劑的差別在於是否產生內生活性。全致效劑內生活性為 1,部分致效劑小於 1。拮抗劑又分競爭性與非競爭性:競爭性結合同一位點、非競爭性結合異位點或堵住通道。 Agonists and antagonists differ in intrinsic activity. A full agonist has an intrinsic activity of 1 and a partial agonist has less. Antagonists are competitive or non-competitive: competitive binds the same site, non-competitive binds elsewhere or blocks the channel.
長期使用致效劑會讓受體下降調節,而下降有兩種方式。去敏感化是磷酸化使受體與 G 蛋白脫鉤,快速可逆且受體數量不變;下調是受體被內吞降解,緩慢且受體數量真的減少。 Chronic agonist use causes downregulation, and there are two forms. Desensitisation is phosphorylation that uncouples the receptor from the G protein, fast and reversible with receptor numbers unchanged. Downregulation is internalisation and degradation, slow, with receptor numbers actually falling.
這是一個雙節點的訊息系統,交感與副交感控制同一批效應器但方向相反。交感負責應付緊急,副交感負責恢復與儲存,腸神經系統則可獨立運作。 This is a two-node signalling system in which sympathetic and parasympathetic control the same effectors in opposite directions. Sympathetic handles emergencies, parasympathetic restores and stores, and the enteric system can run independently.
訊號走三段:節前纖維、神經節、節後纖維再到效應器。關鍵在於節前纖維的傳遞物質固定不變,因為交感與副交感在神經節處無法區分,所以必定使用同一種物質,而節後才分化出系統身份。 The signal passes through three stages: pre-ganglionic fibre, ganglion, post-ganglionic fibre, then effector. Pre-ganglionic transmission is fixed because the two divisions cannot be distinguished at the ganglion, so the same transmitter is used and identity only emerges after the ganglion.
所有節前纖維都是 ACh 作用在菸鹼型受體。節後副交感仍是 ACh 但改用毒蕈鹼型受體,節後交感改用正腎上腺素作用在腎上腺素受體。唯一例外是汗腺,交感節後纖維釋放 ACh 而非 NE。 All pre-ganglionic fibres use acetylcholine on nicotinic receptors. Post-ganglionic parasympathetic fibres keep acetylcholine but switch to muscarinic receptors, while post-ganglionic sympathetic fibres switch to norepinephrine on adrenoceptors. The single exception is the sweat gland, where sympathetic fibres release acetylcholine.
腎上腺素受體的亞型決定一切。α1 走 Gq 升高鈣離子使血管收縮,α2 走 Gi 抑制正腎上腺素釋放,β1 走 Gs 使心跳加快,β2 走 Gs 使平滑肌舒張,β3 走 Gs 使膀胱舒張。毒蕈鹼型中奇數是 Gq、偶數是 Gi,菸鹼型是離子通道所以最快。 The adrenoceptor subtype determines everything. α1 uses Gq to raise calcium and constrict vessels, α2 uses Gi to inhibit norepinephrine release, β1 uses Gs to speed the heart, β2 uses Gs to relax smooth muscle, and β3 uses Gs to relax the bladder. Among muscarinic receptors the odd numbers are Gq and the even numbers are Gi, while nicotinic receptors are ion channels and therefore fastest.
同一條 Gs 鏈在不同效應器上方向相反,在心肌造成收縮、在平滑肌造成舒張。原因是心肌靠鈣離子收縮,平滑肌靠 cAMP 激活 MLCP 移除磷酸基而無法結合。所以平滑肌裡 Gq 收縮、Gs 舒張,兩條鏈方向相反。 The same Gs chain runs in opposite directions in different effectors, contracting cardiac muscle but relaxing smooth muscle. Cardiac muscle contracts because calcium rises, while smooth muscle relaxes because cAMP activates MLCP and removes the phosphate groups needed for binding.
三種介入方式。直接型自己結合受體;間接型提高體內既有的正腎上腺素,三種手法分別是搬出囊泡儲存、阻斷回收、阻斷代謝;混合型兩者兼具。結構上,去掉兒茶酚環的非兒茶酚胺能逃過 COMT 與 MAO,因此可口服且半衰期長。 There are three modes of intervention. Direct agents bind the receptor themselves. Indirect agents raise existing norepinephrine by mobilising vesicular stores, blocking reuptake, or blocking metabolism. Mixed agents do both. Structurally, non-catecholamines lack the catechol ring, escape COMT and MAO, and so become orally viable with a longer half-life.
阻斷的代價來自被關掉的受體原本負責什麼。阻斷 β2 使支氣管無法舒張而痙攣,所以非選擇性 β 阻斷劑在氣喘禁忌。阻斷 α2 移除突觸前負反饋,反而使正腎上腺素釋放增加而血壓反升。 The cost of blockade equals what the blocked receptor normally does. Blocking β2 prevents bronchial relaxation and causes spasm, so non-selective beta blockers are contraindicated in asthma. Blocking α2 removes presynaptic negative feedback, so norepinephrine release increases and blood pressure rises instead of falling.
心血管系統是幫浦加上管路加上調節。血壓等於心輸出乘以全身血管阻力,所以要降血壓只有三條路:降低心輸出、降低血管阻力、或兩者同時做。 The cardiovascular system is a pump, a set of pipes and a regulator. Blood pressure equals cardiac output times systemic vascular resistance, so there are only three ways to lower it: reduce output, reduce resistance, or both.
調節訊號來自三套系統:交感神經、腎素血管張力素醛固酮系統、以及壓力感受器反射。前兩套升高血壓,第三套在血壓過高時反射性減慢心跳,所以高血壓常伴隨心跳減慢。 Three systems deliver regulatory signals: the sympathetic nervous system, the renin-angiotensin-aldosterone system, and the baroreceptor reflex. The first two raise blood pressure while the third reflexively slows the heart when pressure is high, so hypertension often comes with bradycardia.
降血壓藥物有四個作用點。腎上腺素受體作用在血管與心臟,鈣離子通道阻斷劑分血管選擇與心臟選擇兩類,腎素血管張力素系統靠 ACE 與 Ang II 受體,利尿劑作用在腎臟的再吸收位置。 Antihypertensives act at four points. Adrenoceptors act on vessels and heart, calcium channel blockers split into vessel-selective and heart-selective, the renin-angiotensin system is targeted through ACE and the angiotensin receptor, and diuretics act on renal reabsorption sites.
四個槓桿各有用不同的訊號機制。ACEI 阻斷 Ang I 轉成 Ang II 並減少醛固酮。β1 阻斷劑降低心率與腎素。Dihydropyridine 類鈣離子通道阻斷劑阻斷血管的鈣內流。利尿劑阻斷鈉的再吸收而讓鈉從尿中排出。 Each lever uses a different mechanism. ACE inhibitors block the conversion of angiotensin I and reduce aldosterone. Beta-1 blockers lower heart rate and renin. Dihydropyridine calcium channel blockers block calcium entry into vessels. Diuretics block sodium reabsorption so sodium leaves through urine.
兩組藥物的方向相反,而且兩組的鉀離子與鈣離子效果相反。環利尿劑阻斷鈉的同時也阻斷鈣的被動再吸收,所以造成低血鈣;Thiazide 作用處的鈣仍可主動運輸,所以造成高血鈣。非 Dihydropyridine 類鈣離子通道阻斷劑作用在心臟而非血管。 Two pairs of drugs run in opposite directions, one on potassium and one on calcium. Loop diuretics block sodium and with it the passive reabsorption of calcium, causing hypocalcaemia, whereas thiazides leave calcium transport intact and cause hypercalcaemia. Non-dihydropyridine calcium channel blockers act on the heart rather than vessels.
抗心律不整藥物依機轉分為四類,依序阻鈉通道、阻斷 β 受體、阻鉀通道、阻鈣通道。第一類具有使用依賴性,因為它結合鈉通道的不活化狀態,心率越快、不活化通道比例越高,阻斷就越強。 Antiarrhythmics are grouped by mechanism into four classes blocking sodium channels, beta receptors, potassium channels and calcium channels in turn. Class one shows use-dependence because it binds the inactivated state of the sodium channel, so a faster rate means more inactivated channels and stronger blockade.
阻斷的代價包括乾咳與高血鉀。ACE 也分解 Bradykinin,所以抑制 ACE 會讓 Bradykinin 累積而刺激呼吸道產生乾咳;Ang II 下降使醛固酮減少,腎臟排鉀減少而高血鉀。雙側腎動脈狹窄時禁用 ACEI,因為腎灌流壓正靠 Ang II 維持。 The costs of blockade include cough and hyperkalaemia. ACE also degrades bradykinin, so inhibiting it lets bradykinin accumulate and stimulate the airway, causing dry cough. Falling angiotensin II reduces aldosterone, so the kidney excretes less potassium and potassium rises. ACE inhibitors are contraindicated in bilateral renal artery stenosis because glomerular filtration pressure depends on angiotensin II.
呼吸系統有兩件互相獨立的事:氣道張力與氣道發炎。支氣管擴張只處理前者,抗發炎藥處理後者,兩者缺一不可。 The respiratory system has two independent problems: airway tone and airway inflammation. Bronchodilation handles only the first and anti-inflammatory drugs handle the second, so neither alone is sufficient.
支氣管平滑肌同時受交感與副交感支配。交感經 β2 使平滑肌舒張,副交感經 M3 使平滑肌收縮,所以哮喘是這條平衡被打破的結果。 Bronchial smooth muscle is under both sympathetic and parasympathetic control. Sympathetic β2 relaxes it while parasympathetic M3 contracts it, so asthma is a breakdown of that balance.
三類支氣管擴張劑綁三個不同目標:β2 致效劑bind β2 受體,抗膽鹼藥阻斷 M3,Theophylline 抑制磷酸二酯酶以維持 cAMP。另外四類抗發炎藥分別作用在發炎介質的生成、受體、合成酵素、細胞脫顆粒與分解酵素。 Three bronchodilator classes target three different things: beta-2 agonists bind β2 receptors, antimuscarinics block M3, and theophylline inhibits phosphodiesterase to maintain cAMP. Four anti-inflammatory classes act on mediator generation, the receptor, the synthesis enzyme, cell degranulation, and the breakdown enzyme.
β2 走 Gs 升高 cAMP,cAMP 激活 MLCP 讓肌球蛋白輕鏈去磷酸化而無法與肌動蛋白結合,因此平滑肌舒張。支氣管擴張就是這條鏈的結果。 β2 uses Gs to raise cAMP, and cAMP activates MLCP so the myosin light chain loses its phosphate groups and cannot bind actin, which relaxes smooth muscle. Bronchodilation is the result of this chain.
緩解藥與控制藥的方向相反。控制藥(吸入類固醇)作用在生成端,需要誘導新蛋白質合成,所以要數小時至數天;緩解藥(SABA)直接作用在效應端的肌肉上,所以是分鐘級。急性發作時控制藥來不及。 Relievers and controllers work in opposite directions. Controllers such as inhaled steroids act on the generation side and require new protein synthesis, taking hours to days, while relievers such as short-acting beta-2 agonists act directly on the effector muscle within minutes. During an acute attack the controller arrives too late.
兩條治療階梯方向相反。哮喘階梯往上是加抗發炎藥,核心是吸入類固醇;慢性阻塞性肺病階梯往上是加支氣管擴張藥,核心是 LAMA 與 LABA,類固醇效果較差。 The two treatment ladders run in opposite directions. The asthma ladder adds anti-inflammatory therapy upward with inhaled steroids at its core, while the COPD ladder adds bronchodilation upward with LAMA and LABA at its core and steroids work less well.
第一代抗組織胺阻斷 M 受體但脂溶性高能通過血腦屏障,所以產生嗜睡。去充血劑阻斷 α1 讓鼻黏膜血管收縮而消腫,但長期使用使受體下調,產生反彈性充血。第一代抗組織胺的好處是嗜睡可用來治暈車。 The cost of blockade shows in the side effects. First-generation antihistamines block muscarinic receptors but are lipid-soluble enough to cross the blood-brain barrier, so they sedate. Decongestants block α1 to shrink nasal vessels, but long-term use downregulates receptors and causes rebound congestion.
中樞神經系統是一個受血腦屏障保護的控制中心,所以藥物能不能進去是第一步問題。進去之後才談得上作用機制。 The central nervous system is a control centre protected by the blood-brain barrier, so whether a drug can enter is the first question. Only after entry does mechanism matter.
訊號沿著神經元傳遞,最後由神經傳遞物質作用於受體。多數神經傳遞物質是胺類或胜肽,其中許多藥物正是模仿或阻斷這些傳遞物質。 Signals travel along neurons and finally act through neurotransmitters on receptors. Most neurotransmitters are amines or peptides, and many drugs either mimic or block them.
要能進入血腦屏障,藥物必須是小分子量、脂溶性高、且非解離狀態。通過之後才作用於特定受體,例如阿片類的 μ 受體、抗癲癇藥的 GABA 受體或 AMPA 受體、止吐藥的 5-HT₃ 或 NK₁ 受體。 To cross the blood-brain barrier a drug must be small, lipid-soluble and unionised. Only after entry does it act on a specific receptor, such as the mu receptor for opioids, GABA or AMPA receptors for antiepileptics, and serotonin or NK1 receptors for antiemetics.
抗癲癇藥有兩種相反策略。強化 GABA 的藥物增加抑制作用,例如苯巴比妥延長氯離子通道開啟時間、苯二氮平增加開啟頻率;另一種直接掐斷興奮性傳遞,例如 Perampanel 阻斷 AMPA 受體、Ezogabine 開啟鉀通道使細胞超極化。兩者目標都是讓神經元難以達到閾值。 Antiepileptics use two opposite strategies. Those enhancing GABA increase inhibition, with barbiturates lengthening chloride channel opening and benzodiazepines increasing its frequency. The other group cuts off excitatory transmission, with perampanel blocking AMPA receptors and ezogabine opening potassium channels to hyperpolarise the cell. Both aim to make the neuron harder to bring to threshold.
COX-1 與 COX-2 的功能相反。COX-1 處理日常維持工作,例如保護胃黏膜與產生血小板因子;COX-2 只在發炎組織被誘導,腫痛熱就是它造成的。COX-2 選擇性抑制劑胃腸道副作用較少但心血管風險較高。 COX-1 and COX-2 do opposite jobs. COX-1 handles daily housekeeping such as gastric protection and platelet production, while COX-2 is induced only at inflamed tissue and produces swelling, pain and heat. Selective COX-2 inhibitors cause fewer gastrointestinal problems but carry higher cardiovascular risk.
藥物分直接與間接。直接作用於受體,例如阿片類;間接提高體內既有的傳遞物質,例如抑制乙醯膽鹼酯酶讓乙醯膽鹼堆積。非處方藥不能口服的原因也是結構決定的,去掉兒茶酚環才能逃過酵素降解。 Drugs act either directly on receptors, as opioids do, or indirectly by raising existing transmitters, as acetylcholinesterase inhibitors do by letting acetylcholine accumulate. Why some drugs cannot be taken orally is also structural: only removing the catechol ring allows escape from enzymatic degradation.
阻斷的代價從受體分佈推導。阿片類阻斷 μ 受體會同時抑制呼吸中樞而導致呼吸抑制,這是致死主因。長期使用苯二氮平使 GABA 受體下調,產生耐受性與依賴性,所以長期抗焦慮首選 SSRI 而不是苯二氮平。膽鹼酯酶抑制劑只改善症狀,因為阿茲海默的病因是神經元退化。 The cost of blockade follows from where receptors sit. Blocking mu receptors with an opioid also suppresses the respiratory centre, which is the main cause of death. Chronic benzodiazepine use downregulates GABA receptors, producing tolerance and dependence, which is why long-term anxiety treatment uses SSRIs instead. Cholinesterase inhibitors improve symptoms only, because Alzheimer's is caused by neuronal degeneration.
# 跨系統對照:同一個陷阱在六科的不同外觀
這三堂不是受體系統,但七層一樣適用——只是第 3、4 層從「受體與第二傳訊者」換成「干預對象與作用機轉」。這正是框架存在的理由。
臨床試驗條件嚴控、排除特殊族群,所以真實世界才會暴露 ADR。Pharmacovigilance 就是用來補這個盲點的系統。
這是一個上市後的風險管理系統。ADR 是主要死亡原因之一,其中約 60% 是可預防的,所以需要一套主動找訊號的機制,也就是藥物警戒。
This is a post-marketing risk management system. Adverse drug reactions are a leading cause of death and roughly 60 percent of them are preventable, which is why an active signal-finding system called pharmacovigilance exists.
交互作用可以發生在 ADME 的任何一個階段:吸收時被螯合、分佈時被蛋白置換、代謝時被酵素誘導或抑制、排泄時被競爭同一個分泌通道。
Interactions can occur at any ADME stage: chelation during absorption, protein displacement during distribution, enzyme induction or inhibition during metabolism, and competition for tubular secretion during excretion.
綁的是 ADME 四個階段的處理機制本身。葡萄柚抑制 CYP3A4、Rifampicin 誘導肝酶、Aspirin 把 Warfarin 從白蛋白上擠下來、Aspirin 與 Methotrexate 競爭腎小管分泌。
It binds the handling machinery itself at each ADME stage. Grapefruit inhibits CYP3A4, rifampicin induces hepatic enzymes, aspirin displaces warfarin from albumin, and aspirin competes with methotrexate for tubular secretion.
關鍵在游離濃度。蛋白結合的藥物不具活性,只有游離血漿濃度才有效,所以一種藥把另一種藥從蛋白上擠下來時,游離濃度上升,病人就會中毒。
The free fraction decides everything. Protein-bound drug is inactive and only the free plasma concentration is active, so when one drug displaces another from protein the free concentration rises and the patient becomes toxic.
最常記反的是方向:葡萄柚汁是抑制所以讓藥物濃度上升,St. John's Wort 是誘導所以讓濃度下降。這兩個方向反了,整題就錯。
The classic reversal is direction: grapefruit juice inhibits and therefore raises drug levels, whereas St. John's wort induces and therefore lowers them. Getting these two backwards loses the whole question.
介入的原則只有一條:預期利益必須大於有害反應的風險。實務上要用頻率分級來溝通風險,並要求病人主動告知所有用藥,包括處方藥、成藥、補充品與食物。
Intervention rests on one principle: the expected benefit must exceed the risk of harmful reaction. In practice risk is communicated using frequency grading, and the patient must disclose every drug they take, including prescriptions, over-the-counter products, supplements and food.
代價表現在脆弱族群上:孕婦有 Thalidomide 與 Isotretinoin 的致畸風險,兒童用 Aspirin 會引發 Reye's syndrome,G6PD 缺乏者用 Primaquine 會溶血,長者則因多重用藥風險疊加。頻率分級從 very common(≥1/10)排到 very rare(<1/10,000)。
The cost shows up in vulnerable populations: thalidomide and isotretinoin are teratogenic in pregnancy, aspirin triggers Reye's syndrome in children, primaquine causes haemolysis in G6PD deficiency, and the elderly accumulate risk through polypharmacy. Frequency runs from very common at at least 1 in 10 down to very rare at below 1 in 10,000.
Benefit-risk 敘事就是 MSL 的核心工作:KOL 問的從來不是「有沒有副作用」,而是「在誰身上、多常見、能不能監測」。 另一個 BD 經典案例是 side effect 轉 indication:finasteride 變成 Propecia、minoxidil 變成 Rogaine——同一個分子,換一個適應症就是一門新生意。
化療無差別打所有分裂中的細胞;生物藥打明確靶點。兩者的差別是「選擇性」從哪裡來,代價又回到哪裡去。
這是一個「殺快速分裂細胞」的系統,與另一個「打明確分子靶點」的系統並存。前者廣譜、選擇性差;後者精準,但只能打有那個靶點的病人。
Chemotherapy kills every rapidly dividing cell, so it is broad and poorly selective; biologics strike a defined molecular target, so they are precise but only work in patients who carry that target.
細胞毒性藥隨血流到達全身所有分裂中的組織,所以毛髮、腸道、骨髓一起受害。生物藥是蛋白質,口服會被消化,因此只能注射,且怕熱怕菌、需要冷鏈。
Cytotoxic drugs travel in the blood to all dividing tissue, which is why hair, gut and bone marrow are hit together. Biologics are proteins, so they would be digested if swallowed; they must be injected and need a cold chain.
化療綁的是細胞週期的機器本身:DNA、微管、葉酸代謝。生物藥綁的是細胞表面的明確靶點,如 CD20、TNF、IL-6、IL-1、CD25。
Chemotherapy binds the machinery of the cell cycle: DNA, microtubules, folate metabolism. Biologics bind defined surface targets such as CD20, TNF, IL-6, IL-1 and CD25.
化療靠兩種手法讓細胞死:交聯破壞 DNA,或用假原料卡住核酸合成,細胞無法複製就走向凋亡。生物藥結合靶點後,透過抗體依賴性毒殺、補體毒殺或單純切斷生長訊號來達成效果。
Chemotherapy kills through two moves: cross-linking DNA so it cannot be read, or feeding the cell a fake building block so synthesis stalls and the cell undergoes apoptosis. A biologic binds its target and then kills through antibody-dependent cytotoxicity, complement, or simply by cutting the growth signal.
最容易被記反的一點:化療之所以能殺癌細胞,不是因為它認得癌細胞,而是健康細胞會自我修復,癌細胞不會。所以化療利用的是修復能力的差異,不是魔法子彈。另一個反轉是 Nitrosoureas 是唯一能通過血腦屏障的化療藥。
The trap: chemotherapy does not recognise cancer cells. It works because healthy cells can repair themselves and cancer cells largely cannot, so the therapy exploits a difference in repair capacity, not a magic bullet. A second reversal: nitrosoureas are the only chemotherapy class that crosses the blood-brain barrier.
介入方式由細胞週期決定:只在某些期有效的藥叫 cell-cycle-specific,因此給藥時程與輸注時機和劑量一樣重要;all-phase 的藥則任何時期都殺。多藥併用讓不同機制同時打擊,並降低抗藥出現的機會。
Intervention is decided by the cell cycle. Cell-cycle-specific drugs only work in certain phases, so schedule and infusion timing matter as much as dose, whereas all-phase drugs kill at any point. Combination regimes hit the tumour by several mechanisms at once and reduce the chance that resistance emerges.
代價有三層:Neutropenia 是劑量限制毒性;Bleomycin 必須限量因為會造成肺纖維化;一旦產生 acquired resistance 就必須換線。生物藥的代價則是注射給藥、冷鏈、以及每年可能超過 US$500k 的費用。
The cost comes in three layers: neutropenia is the dose-limiting toxicity, bleomycin must be capped because it causes pulmonary fibrosis, and once acquired resistance appears the drug has to be switched. For biologics the cost is injection-only delivery, cold chain, and a price that can exceed US$500k per year.
Resistance 是商業化的引擎:acquired resistance 直接創造 next-line 產品的市場。Biosimilar(38 款,2025-05)是專利懸崖後的價格競爭戰場。 香港 2021 新症 38,462、五大癌佔 57.9%——流行病學就是 market sizing。BD 在做 deal 時問的第一句就是「這個靶點跟現有標準療法差在哪」。
動物是受倫理把關的研究工具,處理原則是 3Rs;而 CAM 的重點只有兩個字——併用還是取代。
這是一套受嚴格規管、設有倫理門檻的研究工具體系,目的是在把治療放進人體之前,先驗證益處與風險。
This is a strictly regulated, ethics-gated research toolkit whose purpose is to verify benefit and risk before a therapy is put into humans.
路徑是先特意繁殖、非寵物的動物,經妥善規劃並取得必要批准後進行實驗;動物結果理想才邀人體志願者進入臨床試驗。
The route begins with purpose-bred animals rather than pets, proceeds only with proper planning and the required approvals, and only invites human volunteers once animal results are favourable.
綁的是模型生物:最常用的是小鼠與大鼠,約佔所有實驗動物的 95%;另外也有斑馬魚、青蛙、果蠅與線蟲。
It binds the model organism: mice and rats dominate at roughly 95 percent of all laboratory animals, supported by zebrafish, frogs, fruit flies and roundworms.
執行的關鍵是變項隔離——控制所有干擾因子、只改變受測因子,才能把結果歸因為因果。人類體型約為小鼠的 3,000 倍,但幾乎所有小鼠基因都與人類基因功能相同,這就是模型成立的原因。
Execution rests on isolating variables: every confounder is controlled and only the tested factor changes, so results can be attributed to cause. A human is about 3,000 times the mass of a mouse, yet almost every mouse gene has a functional human counterpart, which is why the model holds.
最容易混淆的一組方向詞:Complementary 是與常規醫學併用,Alternative 是取代常規治療。另外 3Rs 常被誤解成單純減少用量,實際上 Replacement、Refinement、Reduction 三個方向同時要求更好的設計。
The easiest pair to confuse: complementary means used alongside conventional medicine, alternative means used instead of it. The 3Rs are also misread as simply using fewer animals; replacement, refinement and reduction together demand better experimental design.
介入的方向有兩條:一條是用非動物方法取代,包括 in silico 模型與 FDA 開發的 AnimalGAN 這類 AI;另一條是理解替代醫學本身的機制,例如針灸刺激中樞神經釋放化學物質,適應原作用於 HPA axis。
Intervention runs along two lines: replacing animals with non-animal methods such as in silico models and AI tools like the FDA's AnimalGAN, and understanding the mechanisms claimed by alternative medicine, such as acupuncture stimulating the central nervous system to release chemicals, or adaptogens acting on the HPA axis.
代價是一句話:天然不等於安全。草藥可以傷肝(例如 kava)、與藥物交互、而且在緊急情況下沒有用;加上 FDA 不預先審批 OTC 草本產品,形成證據薄弱與監管缺口並存的風險。
The cost is one sentence: natural does not mean safe. Herbal products can injure the liver, as kava does, interact with medicines, and are useless in an emergency; because the FDA does not pre-approve over-the-counter herbals, weak evidence and a regulatory gap sit side by side.
Translational research 是 MSL 的母語:把基礎發現轉成人體實踐,正是你在跟 PI 和醫院對話時在做的事。3Rs 與 AI(AnimalGAN)減少動物測試,對 CRO / 平台公司是差異化能力;CAM 的證據落差則是商業宣稱的合規地雷。
每一張都是同一個格式:你哪裡踩線 → 正確說法中文 → 英文版可以直接抄。
你不是在背藥理,你是在建立一套「看到機轉就能說出差異」的能力。這正是 technical BD、MSL、medical affairs 在面試時被問的東西。
| Skill | Module | 怎麼對應 |
|---|---|---|
| 辨別 mechanisms of differentiation | L3 / L7 | 同一個適應症裡,兩個分子的差異寫在受體亞型與選擇性上。你能從 receptor → G protein → direction 推出差異,就等於能做 competitive positioning。 |
| 把 mechanism 翻成 value narrative | L1 / L2 | PK/PD 的語言就是 dosing rationale 與 differentiation Claim 的骨幹。ED50、Emax、therapeutic index 是你在簡報裡會真的用到的四個數字。 |
| Risk-benefit 溝通 | L8 | MSL 最核心的日常。KOL 問的不是「有沒有副作用」,是「在誰身上、多常見、能不能監測」。頻率分級與 PV 就是答案的框架。 |
| Translational 語言 | L9 | 把基礎發現轉成人體益處,是 MSL 跟 PI 對話時的共同詞彙。3Rs 與 AI 替代則是平台型公司的差異化能力。 |
| Resistance 與 lifecycle 思維 | L7 | acquired resistance 創造 next-line 市場,biosimilar 是專利懸崖後的價格戰。看懂這條鏈,你看 pipeline 的角度就跟 BD 一樣。 |
| 讀懂 epidemiology = market sizing | L7 / L4 | 香港 2021 新症 38,462、五大癌佔 57.9%;高血壓診斷門檻因指引而異。這些數字在 BD 的世界裡叫 TAM。 |
練習(122 題,含完全沒有藥名的 Direction 分頁):
drill · 122 items
完整版框架(如果你想看 L1–L6 的原始完整寫法):
七層框架 完整中文版 · Seven-Layer Framework (full EN)
講義原始分頁:
上課講義 L5–L9(含 L7–L9 全文) · Site hub